Re-Establishing Solid Food Consumption in a Person With Traumatic Brain Injury and Extended Food Avoidance/Restriction Using a Multicomponent Behavioral Intervention

1 Theoretical and Research Basis for Treatment

Functional oral intake problems frequently occur following traumatic brain injury (TBI) with prevalence of swallowing problems interfering with oral intake occurring in as many as 93% of cases (Howle et al., 2014). Feeding problems may result from brain injury because injury severity, level of consciousness, motor impairments, and cognitive deficits can impede oral feeding (Howle et al., 2014). An impaired swallow may also occur as a complication of medical interventions required during acute care of a brain injury as well. For example, the incidence of dysphagia in individuals who undergo orotracheal intubation following TBI may be as high as 42% (Kwok et al., 2013). In addition, rehabilitation of feeding after a brain injury can be complicated by emerging behavioral challenges (Mayer, 2004). Regardless of causative factors, failure to establish proper oral intake adversely affects recovery, can lead to malnutrition, and increases risk of aspiration (Hansen et al., 2008b; Via & Mechanick, 2013). Furthermore, failure to recover oral feeding diminishes long-term adjustment and quality of life (Kjaersgaard & Kristensen, 2017).

Addressing problematic feeding is a critical part of recovery from a brain injury. The nutritional needs during acute recovery are significantly greater in individuals with TBI (Lee-Anne et al., 2016) and for this reason, enteral support through nasogastric or percutaneous endoscopic gastrostomy (PEG) feeding tube is often necessary to ensure adequate nutrition and prevent health complications. In a multicenter study of acute, inpatient care for individuals who suffered a TBI, Horn et al. (2015) found that 26.5% of TBI patients received enteral nutritional support and patients who received this intervention during acute care showed superior motor and cognitive improvements than patients who did not. However, while enteral feeding support is clearly valuable, return to normal feeding remains problematic. For example, Kim and Suh (2018) found that nearly all TBI patients who underwent removal of nasogastronomy tube feeding showed some degree of dysphagia requiring continued support while 41% never recovered normal swallow. Hansen et al. (2008a) reviewed 173 cases of severe TBI in an inpatient, subacute setting and found that 93% presented with impaired functional oral intake on admission with 36% continuing to require support for oral feeding after 4 months of treatment.

Despite the aforementioned concerns, there is limited research for treatment of behaviorally problematic feeding in adults with TBI. Yuen and Hartwick (1992) described an adult male with TBI and cerebral vascular accident with intact swallow who spit out solid foods after mastication and consumed only pureed foods and nutritional supplementation. Treatment involved fading food texture, appearance, and presentation until the man was able to consume nonmodified foods. A similar approach was replicated with brain-injured children (Dematteo et al., 2002). Hartnedy and Mozzoni (2000) employed an antecedent-based behavioral intervention, finding that oral consumption in brain-injured children improved following reduction of distracting stimuli during meals. Castaño and Capdevila (2010) described four cases in adults with brain injuries where swallow was intact yet unusual behavior disorders complicated normal oral food consumption. Similar to the case report presented herein, one of the adults refused to ingest food orally and another adult displayed odd and episodic food avoidance and refusal. Both behavioral presentations impaired the individual’s ability to establish independent oral food intake despite intact swallow ability, and unfortunately, no effective behavioral treatments were reported.

In the present case report, we describe the behavioral treatment of an adult who had TBI, prolonged food avoidance, and restricted consumption despite adequate swallow ability. Intervention planning followed a clinical-behavioral assessment and an individualized multicomponent behavior intervention was developed combining antecedent and reinforcement strategies. The case illustrates a model of behavioral intervention derived from applied behavior analysis and neurorehabilitation (Heinicke & Carr, 2014; Ylvisaker et al., 2007) and represents one of the few applications with adults referred for oral feeding treatment (Castaño & Capdevila, 2010; Yuen & Hartwick, 1992).

2 Case Presentation

Tammy (a pseudonym) was a 26-year-old woman who suffered a TBI 3 years prior to the start of this study. At that time, she lived in a community-based group home for individuals with acquired brain injuries and received support from care providers at all hours. She also attended a community-based day program with a specialty track for individuals recovering from brain injuries. Her rehabilitation focused primarily on self-care and daily living skills, speech, occupational, and physical therapies, pre-vocational training, and supported visits with her family. In addition, care providers were present with Tammy to support her feeding during daily meals.

Tammy carried several psychiatric diagnoses at the time of the study, specifically generalized anxiety disorder (Diagnostic and Statistical Manual of Mental Disorders [5th ed.; DSM-5; American Psychiatric Association, 2013], 300.02; ICD-10, F41.1) (diagnoses referenced here and throughout: American Psychiatric Association, 2013; World Health Organization, 2018), depressive disorder (DSM-5, 311.0; ICD-10, F32.9), and a history of alcohol abuse (DSM-5 300.05; ICD-10 F10.10). Her food refusal had been continuously present for over 30 months and was characterized as dysphagia unspecified (ICD-10, R1310). However, records indicated that a modified barium swallow study determined that her swallow was adequate for an unmodified diet and a speech pathologist approved a diet of all textures and liquid thicknesses. A nursing assessment and direct behavioral observations concluded that she was able to take small bites and sips without coughing, gagging, choking, or any other concerning signs. Neurofunctional and motor-skeletal observations identified muscle weakness in limbs, abnormal posture, left heal-cord contracture, left hemispatial neglect, and a range of cognitive deficits. Tammy required some degree of direct support for all self-care, continence, ambulation, and activities of daily living.

At the time of this treatment, Tammy was prescribed sertraline (100 mg), divalproex sodium (250 mg), and melatonin (3 mg) daily. She previously received Jevity^® 1.5 Cal (240 cc), 3 times per day, through a PEG tube which had been removed several weeks prior to the study. Food supplementation (Ensure^®) was provided for failure to consume sufficient foods during meals.

3 Presenting Complaints

Behavioral consultation was sought 4 weeks after the removal of the PEG tube when introduction of oral feeding did not improve Tammy’s oral consumption. Prior to the removal of the tube, she would say “I am hungry” and with prompting would swallow liquids and occasionally solid foods, tolerating all textures without signs of dysphagia. She would consume approximately 25% of her foods by mouth, while continuing to receive g-tube feedings 3 times daily. Her primary care provider removed the PEG tube on this basis, assuming an improvement in oral consumption would follow as her appetite was restored.

However, during the weeks after removal of the PEG tube, improvement in oral consumption was insufficient. It was observed that Tammy consumed substantially less than 50% of solid foods by mouth and her cooperation with oral feeding required care provider guidance, prompting, close supervision, and food modifications to achieve only inconsistent consumption. Furthermore, she frequently required food supplementation due to inadequate caloric intake. During oral feeding, she refused to eat solid foods, saying “I’m not hungry,” “I don’t like it,” “I don’t want to” when prompted. She also spit out foods and threw foods into the trash at an opportune moment. This pattern of behavioral challenges matched reports in her record by multiple specialists attempting to re-establish oral consumption in the prior years. Despite continued attempts to improve her feeding, Tammy failed to consume foods sufficient for her caloric needs and was at risk of failure to thrive and gastronomy tube re-implantation.

4 History

Tammy suffered a severe TBI from a motor vehicle accident when she was 23 years old. She was ejected from the vehicle and found at the scene unresponsive and seizing. Her breathing was compromised and she was intubated on site.

Post-accident sequelae included complex right leg fractures, multiple facial fractures, and mediastinal and pulmonary contusions. Tammy was found to have a depressed skull fracture, right side, over the parietal lobe, with cerebral edema and hydrocephalus which eventually required implantation of a ventriculoperitoneal shunt. She spent 5 weeks in a medically induced coma. When her coma was lifted, she was confused, anxious, and unable to talk. As verbal abilities improved, Tammy displayed significant post-traumatic amnesia, was unable to recall immediate events of the accident, lost numerous autobiographical memories, and could not recognize family members. Intensive rehabilitation led to partial gains in mobility, cognition, and self-care.

Tammy exhibited severe challenging behaviors during acute recovery and extended inpatient rehabilitation, notably biting and pinching care providers, destroying her wheelchair accessories, and attempting to unbuckle restraining belts and leave the wheelchair. These behaviors were most often observed during supported self-care activities and feeding. During feeding, she threw and rejected foods. Initially diagnosed with dysphagia, her swallowing greatly improved, and she was able to swallow all liquids and masticate some solid foods with limited supports. However, she spit out foods of all textures, and she often refused foods entirely, consuming less than 75% to 50% of her nutritional needs daily. She was maintained on PEG tube at the time she was discharged to a community-integrated residential program.

Tammy’s condition was further complicated by the presence of paroxysmal movements elicited upon introduction of solid foods and sometimes liquids into her mouth. Movement disorders of this type have been reported in people with TBI (Hannawi et al., 2016). When foods were taken orally, she exhibited a significant facial grimace and twitch, and her left arm would jerk upward sometimes completing a motion to her forehead where she would rest her fingers. Tammy described the movement as involuntary and said it occurred because the food felt cold or “funny” to her. She denied pain but her appearance was of discomfort. The movement was only observed during eating although other odd movements and twitches were seen during bathing. She reported that the sensation was of the water being too hot or too cold even though the temperature was neutral. These perceptions appeared to be sensory processing abnormalities secondary to parietal lobe damage (Mori & Yamadori, 1989).

5 Assessment

Magnetic resonance imaging (MRI) performed 2.5 years after Tammy’s brain injury found evidence of the previous right depressed skull fracture, right parietal damage, and anoxic injuries. Images revealed Wallerian degeneration suspicious for diffuse axonal injuries and bilateral cortical laminar necrosis with right hemisphere significantly greater than left.

Neuropsychological testing conducted at about the time of the study found that Tammy had clinically significant problems with learning and memory on The California Verbal Learning Test 2 (Woods et al., 2006). Additional tests revealed that her receptive language was generally within normal limits but expressive language and reading abilities were greatly diminished. Results of the Boston Naming Test (Goldstein & Beers, 2004) and the Wide Range Achievement Test, Reading (Wilkinson & Robertson, 2006) were below the first percentile. On the Wechsler Adult Intelligence Scale (Weiss et al., 2010), Tammy’s general intellectual abilities were borderline impaired, with Full Scale IQ of 66 (Verbal 66; Performance, 72). She performed at below average or severely impaired in areas of reasoning, sequencing, executive control, and cognitive-processing speed. Additional assessment using the Leiter 3 (Roid & Koch, 2017) found further evidence of deficits in attention, organization, impulse control, as well as poor spatial planning and visuospatial processing.

Functional/adaptive abilities were evaluated using the Mayo-Portland Adaptive Inventory, 4th Edition (MPAI-4) (Malec et al., 2003) and revealed a total MPAI raw score of 56 (T-score = 52) and the following raw scores on MPAI subscales: Ability Scale, 25 (T = 56), Adjustment Scale, 9 (T = 38), and Participation Scale, 22 (T = 52). All scores were indicative of Tammy having moderate to severe limitations compared to national sample of people with ABI receiving services in community-based residential rehabilitation facilities. Her score on the Adjustment subscale indicated relatively less impairment and was supported by direct observation of a generally cooperative nature with her rehabilitation, positive relationships with her family and caregivers, and relatively stable mood with current treatments.

The Overt Behaviour Scale (OBS) (Kelly et al., 2019) was used to evaluate current behavioral concerns that Tammy presented and there were no findings in the area of aggression (physical or verbal), property damage, or sexualized behavior (inappropriate social behavior), although these were observed early in her inpatient rehabilitation.

The authors conducted a functional behavior assessment (FBA) comprising direct observations of attempts to feed Tammy solid foods, interviews with care providers, and review of food intake recordings during meals (Rahman et al., 2013). The supervising clinician (first author) summarized and presented the FBA findings to care providers for clarification and refinement. Key results revealed a range of behaviors that resulted in food avoidance during meals that were described as socially “polite” when contrasted with Tammy’s history of aggression and property damage during meals as reported in the record. For example, when presented with a food item she might verbalize, “No thank you,” “I’m not hungry,” or “I don’t like it.” Prior to presentation of foods she would affirm “I’m hungry” and would select preferred foods from options. If she were cajoled into “trying” a food, she would sometimes swallow, but often spit it out almost immediately. On some occasions, Tammy threw the food into a garbage receptacle.

Furthermore, Tammy appeared uncomfortable during mastication and when she swallowed, showing vigorous twitching, facial grimacing, and stereotyped movement of her left arm although there was no evidence of dental or other oral problems that might evoke this response or cause pain. When asked about her reaction, Tammy said “It feels funny” or “It feels cold” notwithstanding the foods were neutral temperature. It was also observed that she sometimes over-masticated by chewing her foods into a watery paste without swallowing, adding more foods into her mouth before swallowing, or taking larger bites than she could safely ingest. All of these behaviors could be managed by simple prompts such as “take small bites,” “try to swallow now,” and “don’t take another bite until you swallow first.” Finally, Tammy sometimes impulsively grabbed nearby liquids at the start of meals and drank them quickly, consuming a large amount in a small period of time.

A range of positive behaviors were observed as well. First, while Tammy required considerable care provider supervision and support to consume solid foods, she was sometimes compliant and did not resist. Consumption of foods was more likely to occur when care providers sat with her during meals and throwing food away was more likely to occur when they were not by her side. Tammy indicated foods she wanted on her plate from available choices and drank beverages independently. She did not choke or cough when swallowing liquids and solids nor gagged, wretched, or vomited when prompted to consume a particular food despite saying she did not like that food being offered. Furthermore, she usually said “Tastes good” after consuming selected foods she previously stated were not preferred. These observations suggested that food avoidance was not determined by swallowing deficits or taste aversion and that Tammy was highly motivated to please care providers as evidenced by her asking “Am I doing it right?” and “Aren’t you proud of me?” and she enjoyed singing songs to caregivers, joking, and causing them to laugh or smile.

6 Case Conceptualization

We conceptualized Tammy’s feeding problems as a combination of learned operant responses that functioned to avoid solid food consumption and neuropsychological deficits secondary to brain injury.

FBA results confirmed that the challenging behaviors Tammy demonstrated during solid food presentations interfered with oral consumption. Tammy resisted food intake by spitting food from her mouth and throwing food into the trash. She appeared to use polite statements to avoid solid food presentations (e.g., stating, “No thank you. I’m not hungry.”), resulting in solid foods being removed and meals terminated some of the time. However, during the FBA she ate foods she said “she didn’t like” with later comments that the foods “tasted good” suggesting that self-reported taste aversion or lack of hunger was inconsistent with her actual behavior.

Prolonged chewing, inappropriate bite size, poor bite pacing, and attempting to add more food into her mouth before swallowing were interpreted as procedural skill loss, and planning and organization deficits secondary to brain injury (Howle et al., 2014). This was consistent with the findings of neuropsychological testing.

Accordingly, the objectives of assessment-informed treatment were to simultaneously (a) prevent or avoid eliciting interfering behaviors during meals, (b) increase the likelihood of oral consumption, (c) interrupt and redirect food avoidant behaviors, (d) reinforce food consumption, and (e) teach the sequence of proper bite size, mastication, and swallowing.

7 Course of Treatment and Assessment of Progress

Sessions occurred during all meals with a baseline assessment followed by a treatment phase and then a period of follow-up observations when no specific interventions were in place. Across all phases, Tammy was provided a liquid nutritional supplement by mouth anytime she consumed less than 50% of a meal, which she always consumed. A single care provider interacted with Tammy during meals and was responsible for implementing baseline and treatment procedures. The duration of meals did not change throughout the phases and was generally between 20 and 30 min per meal.

During the baseline phase (7 days/1 week), the care provider set a plate with preselected food items and a preferred beverage in front of Tammy and prompted her to “Go ahead and eat.” She often consumed her beverage immediately, in large gulps, and without prompt. The care provider visually supervised while attending to kitchen duties or serving another resident and periodically prompted Tammy to try foods. If Tammy stated she did not like a food, the care provider offered her another item. After each meal, the care provider documented the foods Tammy had consumed by reviewing items left on plate and recording the amount consumed by mouth excluding beverages. For example, if Tammy threw her meal away at the start, the care provider recorded “0% consumed.” If her plate showed 75% of food remaining, the care provider recorded “25% consumed” and so forth. Daily breakfast, lunch, and dinner meals were recorded as the average percentage of daily meals consumed.

The treatment phase (42 days/6 weeks) immediately followed baseline assessment. A multicomponent intervention was employed combining strategies to prevent food avoidance, teach a bite-chew-swallow sequence, and reinforce food consumption. The treatment plan (described below) was explained to Tammy before and throughout all sessions. Because feeding assistance in residential rehabilitation programs is often provided by support staff who have little to no knowledge of swallowing disorders and their treatment (Crary & Groher, 2006), the supervising clinician trained the care providers responsible for interacting with Tammy during meals through verbal and written instructions, demonstration, rehearsal, and performance feedback (Reid, 2017) and written in plain language as easily followed steps increasing the likelihood of consistent implementation (Jarmolowicz et al., 2008). The principal treatment procedures, clinical rationale, and evidence basis were as follows:

Results are presented in Figure 1. During the baseline phase, there was a decreasing trend in oral consumption and an average of 41.4% (range = 25%–50%). Treatment was associated with a gradually increasing trend in oral consumption and phase average of 79.9% (range = 50%–100%). Tammy consumed an average of 91.2% (range = 80%–100%) of foods during the final month of treatment. Tammy demonstrated intermittent episodes of challenging behaviors during sessions but not at a level that interfered with treatment or were lasting. Although we did not measure reduction in behavioral interferences directly during the intervention, direct observations by supervisor and clinician noted corresponding reductions in all behavioral interferences during the course of the intervention and absence during follow-up. Her range of food items consumed also increased during treatment.

OPEN IN VIEWER

Figure 1.

Daily food consumption.

8 Complicating Factors

Several complicating factors were associated with but addressed successfully during treatment. First, due to Tammy’s medical concerns, we consulted nurse professionals on the development of feeding guidelines and her primary physician ordered food supplementation to be used when a criteria for adequate food consumption was not achieved. These measures increased the safety of the intervention in addition to evaluating daily food intake data to avoid any prolonged treatment failure.

Second, oral dyskinesia was immediately elicited when Tammy placed solid foods in her mouth and during mastication. She responded well to reassurances and fortunately the response appeared to habituate with reduced intensity during sessions as observed in similar movement disorders (Wali & Wali, 2018).

Also, care providers required direct supervision to maintain intervention integrity. A residential manager observed feeding sessions periodically and participated in care provider training with the supervising clinician. Continuous oversight was necessary to address occasional misapplication of treatment procedures by care providers and maintain Tammy’s progress and clinical improvement over many weeks.

9 Access and Barriers to Care

The financial support for Tammy’s community-based residential setting was through government insurance within a Medicaid waiver program for persons with acquired brain injury. There were no problems with funding her care and treatment during the course of the study.

10 Follow-Up

We continued to monitor and record oral food consumption for an additional 12 weeks post-treatment. The follow-up results presented in Figure 1 (aggregated in 1-month periods) demonstrated average daily oral consumption from 87.1% to 94.0% of meals consumed. During follow-up, the multiprocedure treatment plan had been eliminated and care provider direct support was no longer required during meals. In addition, the oral dyskinesia that Tammy had displayed during baseline and the early stages of treatment was rarely observed at levels that interfered with consumption. Notably, Tammy was now able to enjoy holiday meals with her family which they attributed to her restored ability to eat and enjoy solid foods.

11 Treatment Implications

This case supports FBA and behavioral intervention to address food avoidance and restriction in an adult with TBI. It highlights the process of designing behavioral interventions with multiple procedures that match controlling variables. The results further support antecedent management and differential reinforcement as “well-established” interventions for behavioral challenges in persons with brain injuries (Heinicke & Carr, 2014; Ylvisaker et al., 2007) including cases of severe food avoidance and restriction.

Clinical assessment revealed many neuropsychological features contributing to Tammy’s food avoidance and restricted intake. She demonstrated a poorly executed sequence of normal eating in which her bite-masticate-swallow sequence was inconsistent and included many “errors” associated with bite size, bite pacing, and swallowing at the proper time. The loss of formerly routinized skills following severe TBI was also apparent with Tammy. Accordingly, her treatment required direct skills training during meals using errorless learning in the form of in vivo prompts and corrections (Clare & Jones, 2008). Further relevance of a neuropathological contribution to her clinical presentation is research by Mori and Yamadori (1989) who reported four cases of isolated stroke to the parietal lobe which produced food avoidance and rejecting behaviors despite an intact swallow in some survivors. In addition, neuropsychological testing with Tammy supported the possible contributions of executive dysfunction in the areas of attention, planning, and execution in her skill deficits, all factors that had to be addressed through a function-based treatment plan.

However, skills training alone would have been insufficient as Tammy displayed a long-standing pattern of food avoidance and restricted intake. Several behaviors functioned to avoid solid food from entering her mouth, specifically verbalizing refusal, spitting out food, and throwing food away. In particular, Tammy’s attempts to recruit social reinforcement (e.g., “Did I do a good job? Are you proud of me?”) suggested that copious social reinforcement during feeding sessions might compete with escape motivation.

The treatment implication from Tammy’s long-standing history of aggression and property damage during meals was to minimize feeding demands (Matson et al., 2005) and emphasize prevention-focused antecedent procedures. Her treatment plan emphasized increased social opportunities throughout meals, rapport-building strategies, choice of food items, and prompting short breaks or alternative food items during feeding sessions to maintain a positive experience during feeding sessions and a low demand-to-reinforcement ratio mitigating the need for escape-induced aggression (Green et al., 2005). Of course, these and other procedures were implemented simultaneously so it is not possible to isolate how they separately influenced treatment outcome.

Tammy’s treatment plan was written in plain English and described in a single-page procedural guide to promote intervention integrity among several care providers who supported Tammy during meals (Jarmolowicz et al., 2008). As noted previously, a residential manager was on site to observe meals and provide performance feedback to care providers throughout the course of treatment and in concert with clinician direction as well. However, we did not conduct formal assessment of intervention integrity (Sanetti & Kratochwill, 2014) or assess interobserver agreement (IOA) of data recording (Carter & Wheeler, 2019). While her support staff were generally excited about the intervention and the progress she was making, a formal assessment of treatment acceptability would have enhanced the case study (Sterling-Turner & Watson, 2002). Concerning social validity (Carter & Wheeler, 2019), it is meaningful to point out that Tammy’s father was a professional chef who considered her “ability to eat food again” as one of the most important goals to be achieved at the time of admission to the current setting and an objective that was fully realized at follow-up.

This case also raises issues about the diagnosis of avoidant/restrictive food intake disorder (ARFID, DSM-5 F50.82) (American Psychiatric Association, 2013) and the efficacy of behavioral interventions for treating this condition. Although originally proposed as a childhood psychiatric disorder, ARFID is now recognized to occur in adults (Katzman et al., 2019) and responds effectively to cognitive-behavioral therapy (CBT) (Görmez et al., 2018). Tammy did not receive the diagnosis of ARFID at the time this treatment began but her clinical presentation shared the essential features of resisting oral food consumption that causes insufficient nutritional intake requiring enteral supplementation. Other diagnostic considerations were the lack of evidence she was simply a “picky eater” and no competing psychiatric conditions such as anorexia nervosa which might account for her behavior. As well, the severity of her food avoidance was not a typical of TBI.

With regard to the behavioral treatment of ARFID, Görmez et al. (2018) successfully intervened with an adult using hierarchical exposure to food types, relaxation training, and cognitive restructuring. Roth et al. (2010) treated food refusal in a child with Asperger’s disorder by incorporating choice of foods, fading size and effort demands, and incorporating competing reinforcement. In the case of an adult with ARFID, emetophobia, and other co-occurring psychiatric diagnoses, Rigby (2018) combined patient education about her condition, in vivo and imaginal exposure, and guidance for oral feeding. Although these studies and the present case report featured different procedures, it appears that engaging the patient fully in the treatment process and carefully structuring sessions to slowly increase oral consumption while avoiding food refusal are core strategies contributing to positive outcome. Engaging a person with brain injury in their behavioral treatment conceptualization and intervention design is consistent with other research using an integrated person-centered therapy approach to applied behavioral intervention (Ricciardi et al., 2020).

12 Recommendations to Clinicians and Students

This case demonstrates the importance of functional assessment to identify antecedent and consequence conditions associated with food avoidance and refusal including cognitive and medical influences. In most cases, a function-based treatment plan will include multiple procedures that address the variables responsible for presenting problems. We recommend that clinicians and students reference the research literature to extract evidence-based methods that have been shown to reduce and eliminate food avoidance and refusal in both children and adults with varied clinical conditions (Volkert et al., 2016).

Another feature of this case was the combination of neuropsychology and behavior analysis with both disciplines leading to an integrated case conceptualization that accounted for how deficits caused by a TBI and operant learning contributed to food refusal and avoidance. Hence, clinicians and students should be cognizant of the benefits from multidisciplinary collaboration and case formulation when intervening with persons who have protracted, complex, and previously treatment-resistant feeding problems.