Secondary Logo

Journal Logo

Original Articles: Hepatology and Nutrition

Nutritional and Psychosocial Outcomes of Gastrostomy Tube–Dependent Children Completing an Intensive Inpatient Behavioral Treatment Program

Silverman, Alan H.*; Kirby, Midge*; Clifford, Lisa M.*; Fischer, Elizabeth; Berlin, Kristoffer S.; Rudolph, Colin D.*; Noel, Richard J.*

Author Information
Journal of Pediatric Gastroenterology and Nutrition: November 2013 - Volume 57 - Issue 5 - p 668-672
doi: 10.1097/MPG.0b013e3182a027a3
  • Free


Gastrostomy tube (GT) feeding is used when a child has a feeding problem that is serious enough to compromise his or her nutritional status and growth. Although GT feeding can have important immediate benefits (1,2), the ultimate goal in treatment is to introduce or reintroduce oral feeding.

In several studies, (1,3–6) behavior-based treatments have been shown to be effective in weaning GT-dependent children; however, these studies have been small in size and have not provided information about the long-term outcomes of these treatments. As an example, Byers et al (1) reported outcomes for 9 children 1.8 to 5.5 years of age who completed inpatient care to wean them from GT dependence. Results showed that at the time of discharge, 44% (n = 4) patients had completely discontinued GT feedings, and between the 2- and 4-month follow-up, an additional 2 patients were consuming all of their calories orally. Although promising, this study was limited by the small sample size, behavioral outcomes were not reported, and no longitudinal data were presented.

The primary purpose of this study was to describe an intensive program of behavior therapy designed to wean children who have failed traditional outpatient treatment from dependence on GT feedings.


This was a retrospective study of prospectively and retrospectively collected data associated with a clinical cohort of 77 children diagnosed as having a feeding disorder, GT feeding dependence (>1 year), and an inability to maintain acceptable growth via oral feeding completing an inpatient tube weaning protocol. Children and their families were recruited from the feeding and swallowing center at the Children's Hospital of Wisconsin between January 2005 and December 2009. The children recruited had demonstrated a sustained inability to maintain acceptable growth without supplemental calories from GT feeding. This study was approved by the hospital institutional review board. Written consent from parents and verbal assent from children (as appropriate) were obtained.

Demographic and program data descriptive of the sample are presented in Table 1. Patients’ psychosocial status was assessed at the intake assessment, pretreatment, and posttreatment. Nutritional status was assessed at the intake assessment, pretreatment, posttreatment, 1 month, 3 months, 6 months, and 1 year after treatment.

Demographic and baseline nutritional status of study participants

Children were enrolled into the feeding tube wean protocol based on readiness for oral feeding. Before admission, each child was examined by a team of feeding specialists, including a pediatric gastroenterologist, a dietitian, a speech-language pathologist, and a pediatric psychologist. Each of these families had attempted outpatient therapy before beginning the inpatient treatment program. Children were enrolled in the feeding tube protocol only if they had participated in our outpatient program and had not demonstrated clinical progress (as measured by a reduction in percentage of GT feeding calories) for ≥3 months. Only 10% of the children met this criterion in our outpatient program. Children in the present study are thus representative of the most difficult-to-treat cases when outpatient treatment is attempted.

During the inpatient program, children received treatment from a pediatric psychologist at each meal 3 times per day, 7 days per week, until discharged. At least 1 parent was required to be present at all mealtimes. A dietitian worked closely with the treating psychologist to monitor energy intake (calculated daily), weight, and hydration status. Supplemental feeding was discontinued (or significantly reduced as medically indicated) during the inpatient program to promote hunger; however, additional fluids (water or Pedialyte [Abbott Nutrition, Abbott Park, IL]) were given via tube to prevent dehydration as needed (urine-specific gravity ≥1.020). To ensure treatment fidelity, patients only received the behavioral and nutrition intervention during the time of their admission to the hospital.

Treatment was divided into 3 distinct phases. During the initial phase of treatment, all meals were fed by a team of 4 psychologists who alternated sessions. Caregivers observed the meals via closed circuit television. When possible, another psychologist or a psychology graduate student would sit with the family observing the meal to provide additional teaching about the behavioral techniques that were being modeled. All of the sessions were video recorded to ensure treatment fidelity and to provide the opportunity to review sessions with families as part of their education and training. Behavioral treatments combined well-established techniques to achieve weaning, including environmental controls (7), appetite manipulation (6), contingency contacting (8), re-presentation (9), texture fading (10), and differential reinforcement of other behavior (11). Typically, reinforcement consisted of immediate verbal praise of targeted feeding behaviors, including bite acceptance, chewing, and swallowing, as well as other desirable behaviors, such as self-feeding and use of appropriate table manners. For some children, brief timeouts at the table (feeder turns away from child for approximately 30 seconds) were given when differential reinforcement was insufficient to extinguish an undesirable behavior (eg, tantrums, bite refusal). During the second phase of treatment, caregivers were transitioned into the feeding environment, gradually assuming the role of feeder, whereas the psychologist gradually fades from the feeding environment. In the final phase, psychologists coached caregivers remotely via an earpiece speaker to allow the feeding relation to solidify between caregiver and child in the absence of the psychologist (12). Children were discharged from the hospital when the parents were consistently able to demonstrate ability to maintain the feeding intervention. Outpatient follow-up was provided because many children and families experience some level of behavioral difficulty as they transition home. Behavioral follow-up sessions, often conducted in conjunction with a dietitian, provided close monitoring of behavioral interventions, maintenance of behavioral goals, and energy intake and growth. Follow-up sessions with physicians and speech-language pathologists were scheduled as needed. Follow-up sessions are tapered off as the children become nutritionally stable and behavioral problems in relation to oral feeding become minimal or nonexistent. Measures for analysis included About Your Child's Eating, the Mealtime Behavior Questionnaire, and the Parenting Stress Index Short Form (13).

About Your Child's Eating

The About Your Child's Health (AYCE) is a valid and reliable 25-item parent report instrument that measures parent beliefs and concerns regarding their child's eating, in the areas of frequency of child eating behaviors, parental mealtime interactions with the child, and parent's feelings about mealtimes. The AYCE factor structure consists of 3 scales: Child Resistance to Eating, Positive Mealtime Environment, and Parent Aversion to Mealtime. The clinical cutoff for the 3 scales are 2 standard deviations above (or below) the mean of the normative sample (Child Resistance to Eating >33; Positive Mealtime Environment <12; Parent Aversion to Mealtime >12).

Mealtime Behavior Questionnaire

The Mealtime Behavior Questionnaire has 4 subscales representing various typographies of mealtime problems (Food Manipulation, Distraction/Avoidance, Aggression, and Choking/Gagging/Vomiting) and a total scale score on this measure (14). In this study, the total score is reported. The clinical cutoff for the total scale score is 1.5 standard deviations above the mean score of the normative sample (total score >77).

Parenting Stress Index Short Form

The Parenting Stress Index Short Form measures parent perception of caregiving tasks that are stressful to parents and the extent to which parents underreport these stressors (ie, a validity scale) (15). In this study, the total score is reported. The clinical cutoff for the total scale score is ≥85th percentile (total score >90).

The recorded anthropometric assessment included percentage ideal body weight for height estimated by use of the McLaren method (16); the treatment goal for total energy intake appropriate for age (energy intake only), sex, and adjusted activity level; and oral intake as percentage of caloric goal. It should be noted that although calorie intake refers only to energy intake, from a clinical management perspective, the effort was always to meet major nutrient requirements with food or the use of nutritionally complete formula as tube feeding was decreased. For most of these patients, the introduction of oral feeding usually begins with pureed foods, smooth foods, and formulas with a gradual advancement to chewable foods.

Nonmutually exclusive etiologic biobehavioral classifications were made by the multidisciplinary feeding team following an existing scheme (17), with additional biobehavioral classifications from Crist and Napier-Phillips (18). Classifications included structural/anatomic, neurodevelopmental, cardio/cardio-respiratory, gastrointestinal, metabolic, allergy/immune, endocrine, behavioral, and sensory problems. Previous uses of this system have yield acceptable overall levels of interrater reliability (19).


Seventy-seven enrollees completed the inpatient program and were eligible for analysis based on eligibility criteria (Fig. 1). Demographics and tube feeding data are shown in Table 1. The patients were classified according to medical and psychosocial comorbidities. A majority of patients had gastrointestinal (92%), behavioral (90%), neurodevelopmental (68%), or cardiorespiratory (51%) comorbidities. Other comorbidities were classified as structural upper-airway anomaly (22%), allergic or immunodeficiency (17%), sensory (16%), endocrine (8%), or metabolic (5%).

Patient enrollment.

Percentage of ideal body weight was 98% ± 0.8% at hospital admission and 96% ± 0.7% (mean ± standard error of the mean [SEM]) at discharge. The mean duration of hospitalization was 10.9 days (standard deviation [SD] 2.04), over which oral intake percentage increased from 30% ± 2.5% to 82% ± 3% (mean ± SEM). The increase in caloric intake, measured in terms of percentage of goal, was sustained during the 12-month follow-up period, being equal to 85% ± 3.6% (mean ± SEM) at 12 months (Fig. 2). It is noteworthy that patients tended to lose weight during the initial phase of treatment and when caregivers transitioned to the role of feeder; however, children typically gained weight well with caregivers feeding before being discharged from the hospital. Average weight loss from admission to lowest weight during the program was 4.0% (0%–11%; SD 2.5%).

Longitudinal nutritional status of study participants.

Patients averaged only 13.3% oral calories (range 0%–58%; SD 21.1%) at the initial outpatient appointment. From pre-treatment to admission, oral calorie intake improvement was observed in many patients during the period of time when they were being treated in the outpatient clinic; however, if they showed no progress in reaching their treatment goals in 30 days (≥2 consecutive appointments), they became eligible for admission to the inpatient program. At admission, patients obtained by oral feeding an average of 28% (range 0%–113%; SD 23%) of their caloric goal. By the end of the inpatient period, patients obtained by oral feeding 83% (range 12%–167%; SD 27.8%) of their caloric goal. At discharge, 51% of patients needed no GT feeding, and 1 year after discharge, an additional 12% of patients needed no GT feeding in the follow-up outpatient clinic within the first year following hospital discharge.

Weight loss was noted for most patients pre- to post-treatment and continuing up to 1 month from discharge. The trend typically reversed by 3 months after discharge, with children returning to baseline percentage ideal body weight by 1 year from discharge. Despite a pattern of weight loss, patients typically did not lose enough weight to be considered at medical risk (>5% and/or body weight <90% ideal) (Fig. 1). Patients unable to complete total weaning from GT feeding also reduced GT dependence. Of those patients unable to be completely weaned from GT dependence, 3 patients reduced tube dependence by up to 25%; 2 patients had a 25% to 50% reduction in tube energy; 4 patients had a 50% to 75% reduction in tube calories; and 11 patients had reduced tube dependence of >75%, but continued to receive a relatively small percentage of nutrition from tube feeding.

Regarding actual removal of gastrostomy tubes, at the time of discharge from our interdisciplinary program, 14 patients had tubes removed, given that they took in orally all required calories, fluid, and medications. Forty-seven patients still had gastrostomy tubes for fluid, medications, or a fraction of calories. Data on the remaining 11 patients are not available because the patients were unreachable.

Data regarding psychosocial treatment effects are presented in Table 2. These data show changes, which occurred from the start of the inpatient admission to the immediate postassessment after discharge from the hospital. Overall, interactions between caregivers and children were described more positively as measured by the AYCE. Most notably, Child Resistance to Eating (eg, “My child is a picky eater,” “I have to force my child to eat.”) was significantly reduced (T = 51.22, P < 0.000). Likewise, Parent Aversion to Mealtime (eg, “I dread mealtimes,” “It is hard for me to eat with my child.”) was significantly improved (F = 18.76, P < 0.000). Finally, Positive Mealtime Environment (eg, “The family looks forward to meals together,” “I get pleasure from watching my child eating.”) was significantly improved (F = 9.13, P = 0.004). Directly observable problem mealtime behaviors (eg, food manipulation, distraction/avoidance, aggression, choking/gagging/vomiting) were also significantly improved as measured by the Mealtime Behavior Questionnaire (F = 60.91, P < 0.0001). Finally, parenting-related stress, as measured by the Parenting Stress Index, was reduced but was not found to be statistically significant (F = 3.11, P = 0.086).

Paired samples tests for psychological effects


Our data demonstrate that medically complex tube-fed children who make no sustained progress toward feeding tube independence when treated in an outpatient setting can be completely or partially weaned off tube feeding when treated in an inpatient setting through the intense application of behavioral strategies without detriment to their health or nutritional status. The participants not only exhibited extreme resistance to oral feeding but they also experienced a variety of comorbid conditions, including gastrointestinal disorders (eg, reflux, vomiting, omphalocoele, esophageal atresia, congenital diaphragmatic hernia), neurodevelopmental disorders (eg, seizure disorders, cerebral palsy, intracerebral hemorrhage), and/or cardiorespiratory disorders (eg, congenital heart disease, bronchopulmonary dysplasia, pulmonary lobular resection) that prevented the establishment of normal feeding. These comorbid medical conditions improved over time because of medical therapies, operative procedures, or further growth with neuromuscular development, such that the feeding disorder and tube feeding dependence was the prominent feature limiting normal social interaction and activities of daily living.

Intake data demonstrate that although these children had grown adequately via tube feeding, their oral caloric intake was minimal, averaging 14% of caloric goal, and psychosocial instruments documented mealtime behavior problems with significant resistance to oral feeding goal. They also exhibited, according to parent report and our own observations, a variety of extreme behaviors when they fed orally. These behaviors included tantrums, refusal to swallow contents of the mouth, vomiting, or physical manifestations of anger or fear within this population of children. Understandably, parents were adversely affected by the stress related to feeding their children, which contributed to parents developing an aversion to mealtimes.

Outpatient therapy is sufficient to transition the majority of tube-fed children to feeding tube independence. The children in our study are representative of those children who fail to make progress when treated as outpatients; however, as demonstrated by the present results, even those difficult-to-treat children can make significant progress toward feeding tube independence when treated in an inpatient setting where they are closely monitored and treated with an intense regimen of behavioral techniques.

The inpatient treatment increased oral calories to a mean of 82% of the caloric goal in just 11 days. The hospital setting allows medically safe reductions or elimination of tube feedings to maximize appetite while carefully monitoring weight, hydration, and glycemia/ketosis; however, when initially hospitalized, the participants lost weight, albeit not to a medically significant degree (average weight loss 2.0% of ideal body weight during the hospital admission). Weight loss was regulated by using supplemental tube feeding in 35.1% of participants according to parameters in the clinical protocol. More important, treatment-associated weight loss was recuperated while sustaining improved oral intake (Fig. 2).

Appetite manipulation has been well characterized as a requirement for sustained transition to oral feeding and has been accomplished thorough energy restriction (20) or medication (5). In our patient population, tube feedings were discontinued at inpatient admission and 3 structured meals provided the only opportunity for the child to take in energy and satiate hunger. Our study adds to the existing literature by demonstrating that the use of appetite manipulation in a population of medically complex children is safe provided the treatment is completed in a medical setting in which adequate medical monitoring and oversight are made available.

Our successful outcomes likely reflect the controlled hospital environment, as well as the collective strength of the interdisciplinary approach. The participants were selected for the program by both failure of standard outpatient approaches and a thorough understanding of their abilities by the physician and speech-language pathologist members of the team. Once in the program, the structured environment and targeted behavioral interventions produced results commensurate with the participant's previously delineated abilities, as the psychologists worked with the reassurance that goals were realistic and attainable. It is likely that any other successful approaches to feeding disorders in medically complex patients will have a similar interdisciplinary structure.

Providers who become involved in the treatment of a child with feeding disorder should be aware that the etiological bases of feeding problems are typically organic; however, behavioral factors generally play a prominent role in the maintenance of feeding problems (12). The interplay between medical and behavioral issues is complex and challenges efforts to wean children from tube dependence. Among GT-dependent children, behavioral feeding problems commonly involve avoidance behaviors such as food refusal and tantrums (21,22) secondary to a history of aversive feeding experiences (eg, gagging, choking, painful swallowing) (23). In addition, the steady and consistent delivery of nutritional needs delivered via tube feedings minimizes feelings of hunger, thus diminishing the natural motivation for oral intake (24).

Limitations of our study include retrospective data collection and incomplete ascertainment of follow-up data resulting in decreasing sample size through 12 months of follow-up analysis. The data loss can be attributed to patients being discharged from our clinic, missing visits, or coming to visits with missing or incomplete dietary records. Heterogeneity of the patient population may also be construed as a limitation; however, given the strong treatment effects across a broad range of patient etiologies, it may also represent the robust nature of the treatment approach. Finally, psychological measures are dependent upon parent report. It is possible that caregivers may have been influenced by the pre-post nature of the study design and subject to systematic bias in response.

In conclusion, our data demonstrate that inpatient behavioral interventions are highly effective and safe for transitioning chronically tube-fed children to oral feeding. Our data suggest that a combination of nutrition and behavioral interventions is effective for the rapid weaning of children who have failed traditional treatment approaches. Treatment effects are shown to benefit both the child and the adult caregivers as demonstrated by an increase in positive mealtime behaviors as well as improved caregiver–child interactions within the meal context. Furthermore, the outcome is sustainable and independent of the medical complexity of our patients. Future research may establish unique considerations matching specific nutrition and behavioral techniques to the underlying comorbidity (eg, congenital heart disease), which may further optimize transition to oral feeding.


1. Byars KC, Burklow KA, Ferguson K, et al. A multicomponent behavioral program for oral aversion in children dependent on gastrostomy feedings. J Pediatr Gastroenterol Nutr 2003; 37:473–480.
2. Corwin DS, Isaacs JS, Georgeson KE, et al. Weight and length increases in children after gastrostomy placement. J Am Diet Assoc 1996; 96:874–879.
3. Benoit D, Wang EE, Zlotkin SH. Discontinuation of enterostomy tube feeding by behavioral treatment in early childhood: a randomized controlled trial. J Pediatr 2000; 137:498–503.
4. Burmucic K, Trabi T, Deutschmann A, et al. Tube weaning according to the Graz model in two children with Alagille syndrome. Pediatr Transplant 2006; 10:934–937.
5. Davis AM, Bruce AS, Mangiaracina C, et al. Moving from tube to oral feeding in medically fragile nonverbal toddlers. J Pediatr Gastroenterol Nutr 2009; 49:233–236.
6. Kindermann A, Kneepkens CM, Stok A, et al. Discontinuation of tube feeding in young children by hunger provocation. J Pediatr Gastroenterol Nutr 2008; 47:87–91.
7. Silverman AH, Tarbell S. Roberts MC, Steels RG. Pediatric feeding problems, undernutrition, and vomiting disorders. Handbook of Pediatric Psychology 4th ed.New York:Guilford Press; 2009. 425–429.
8. Hoch T, Babbitt RL, Coe DA, et al. Contingency contacting. Combining positive reinforcement and escape extinction procedures to treat persistent food refusal. Behav Modif 1994; 18:106–128.
9. Coe DA, Babbitt RL, Williams KE, et al. Use of extinction and reinforcement to increase food consumption and reduce expulsion. J Appl Behav Anal 1997; 30:581–583.
10. Shore BA, Babbitt RL, Williams KE, et al. Use of texture fading in the treatment of food selectivity. J Appl Behav Anal 1998; 31:621–633.
11. Patel MR, Piazza CC, Martinez CJ, et al. An evaluation of two differential reinforcement procedures with escape extinction to treat food refusal. J Appl Behav Anal 2002; 35:363–374.
12. Linscheid TR. Behavioral treatments for pediatric feeding disorders. Behav Modif 2006; 30:6–23.
13. Davies WH, Ackerman LK, Davies CM, et al. About Your Child's Eating: factor structure and psychometric properties of a feeding relationship measure. Eat Behav 2007; 8:457–463.
14. Berlin KS, Davies W, Silverman AH, et al. Assessing Family-based Feeding Strategies, Strengths, and Mealtime Structure with the Feeding Strategies Questionnaire. J Pediatr Psychol 2009; 36:586–595.
15. Abidin R. The Parenting Stress Index—Short Form. Odessa, FL:Psychological Assessment Resources; 1995.
16. Phillips S, Edlbeck A, Kirby M, et al. Ideal body weight in children. Nutr Clin Pract 2007; 22:240–245.
17. Burklow KA, Phelps AN, Schultz JR, et al. Classifying complex pediatric feeding disorders. J Pediatr Gastroenterol Nutr 1998; 27:143–147.
18. Crist W, Napier-Phillips A. Mealtime behaviors of young children: a comparison of normative and clinical data. J Dev Behav Pediatr 2001; 22:279–286.
19. Berlin KS, Lobato DJ, Pinkos B, et al.Pediatric feeding problems: relationships of nutrition, weight, and behavior to patterns of medical and developmental comorbidities. Paper presented at the meeting of the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition; November 12–14, 2009; National Harbor, MD [abstract no. 88].
20. Kindermann A, Kneepkens CMF, Stok A, et al. Discontinuation of tube feeding in young children by hunger provocation. J Pediatr Gastroenterol Nutr 2008; 47:87–91.
21. Babbitt RL, Hoch TA, Coe DA, et al. Behavioral assessment and treatment of pediatric feeding disorders. J Dev Behav Pediatr 1994; 15:278–291.
22. Silverman AH. Interdisciplinary care for feeding problems in children. Nutr Clin Pract 2010; 25:160–165.
23. Piazza CC, Carroll-Hernandez TA. Tremblay RE, Barr RG, Peters RV. Assessment and treatment of pediatric feeding disorders. Centre of Excellence for Early Development, Encyclopedia on Early Childhood Development. Montreal, Quebec:2004.
24. Linscheid T, Budd K, Rasnake L. Roberts MC. Pediatric feeding problems. Handbook of Pediatric Psychology 3rd ed.New York:Guilford Press; 2003. 481–498.

behavior therapy; clinical outcomes; nutrition; tube weaning

© 2013 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology,