Share this article on:

Role Reversal Method for Treatment of Food Refusal Associated With Infantile Feeding Disorders

Segal, Idit*; Tirosh, Anat; Sinai, Tali; Alony, Sari; Levi, Anat*; Korenfeld, Lia*; Zangen, Tsili*; Mizrachi, Avi*; Boaz, Mona; Levine, Arie*

Journal of Pediatric Gastroenterology and Nutrition: June 2014 - Volume 58 - Issue 6 - p 739–742
doi: 10.1097/MPG.0000000000000309
Original Articles: Hepatology and Nutrition

Objectives: Infantile feeding disorders (IFDs) are common causes of food refusal and failure to thrive, and are frequently encountered by primary care physicians and specialists. We have published the Wolfson criteria for IFD, which have eased the approach to the diagnosis of IFDs. Along with and complementary to the Wolfson criteria, we have also developed the role reversal treatment method for IFD, which has been briefly described earlier. The aim of this study was to validate the role reversal treatment method on a cohort of infants diagnosed as having IFD and to present a detailed description of this method for the first time.

Methods: Parents of infants and children diagnosed as having IFD were invited to participate in the study; they were handed over a questionnaire comprising 6 categories of questions related to patient and parents behaviors, attitudes, and perceptions, which was completed at initiation and at the end of treatment. Full response was defined as improved normative feeding, cessation of abnormal parental feeding, and improved or normal growth patterns. A partial response was defined as success with two-third categories.

Results: We enrolled 38 patients, and 32 patients completed the study. Improved feeding occurred in 78%, full recovery was documented in 53% of infants by 6 months, and partial response was observed in another 25%. All forms of pathological feeding improved significantly (mechanistic, nocturnal, persecutory, forced feeding, and distraction).

Conclusions: The role reversal treatment method is a simple and effective approach to the treatment of food refusal associated with IFD.

Supplemental Digital Content is available in the text

*Pediatric Gastroenterology and Nutrition Unit, Wolfson Medical Center, Holon

School of Nutritional Sciences, Hebrew University, Rehovot

Epidemiology Unit, Wolfson Medical, Center, Tel Aviv University, Israel.

Address correspondence and reprint requests to Idit Segal, MD, Pediatric Gastroenterology and Nutrition Unit, Wolfson Medical Center, PO Box 5, Holon, Israel (e-mail:

Received 7 January, 2014

Accepted 7 January, 2014

I.S., A.T., T.S., and A.L. are senior authors.

Supplemental digital content is available for this article. Direct URL citations appear in the printed text, and links to the digital files are provided in the HTML text of this article on the journal's Web site (

The authors report no conflicts of interest.

Infantile feeding disorders (IFDs) are the leading causes of failure to thrive in infants and toddlers (1–4). The identification and treatment of medical conditions leading to feeding difficulties such as gastroesophageal reflux, food allergy, or swallowing disorders commonly lead to a targeted treatment and subsequent improvement or resolution of symptoms (4–9). Multiple screening (Wolfson criteria) and diagnostic criteria (Wolfson, Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Chatoor, DC: 0–3) are suggested for the diagnosis of IFDs (10–14). Our published data have eased the process of the diagnosis of IFD, which is based on clinical criteria to identify and discriminate functional IFDs from medical causes of feeding refusal (Table 1) (11). We have also observed and confirmed with a prospective study that nearly all IFDs develop through a common pathway involving food refusal and an attempt to circumvent refusal or poor feeding by the use of pathological feeding methods (Table 2) by parents or caregivers (called the intrusive feeding cycle), first proposed by Levine and coworkers (10,11). This study demonstrated that pathological feeding can be identified in 100% of children with IFDs, and provided data from a treatment method we have used (previously called avoidance transfer, now called role reversal therapy) that is geared to abolishing pathological feeding and restoring feeding autonomy (online-only appendix, In a previous study and clinical experience, role reversal therapy led to the resolution of food refusal and restoration of growth in up to 70% of patients <6 years of age, in both retrospective and prospective studies, in which treatment was performed by a small experienced team. Treatment results, however, were not the primary outcomes of these studies, and the ability of pediatricians and gastroenterologists to use this system was not evaluated. We present our results of a prospective validation cohort performed by a team trained to use the role reversal method, and for the first time present a detailed description of this method.





Back to Top | Article Outline


This was a prospective study involving infants ages 2 months to 6 years. The study was approved by a hospital ethical committee, and all of the parents signed an informed consent. Patients referred to a tertiary failure-to-thrive and infant feeding clinic were screened by the earlier described Wolfson criteria (10). The data were collected by a single dietitian (A.T.) between 2011 and 2012 as part of her thesis of a research project on infant and parental responses (eg, behavioral, nutritional, anthropometric) to therapy of IFDs. In the present study, we confine ourselves to the outcomes related to feeding behavior and growth.

All of the patients were screened by a pediatric gastroenterologist or a pediatrician involved in our enteral nutrition program. Patients who had a clear medical reason for food refusal (milk allergy, celiac disease, and dysphagia suggestive of eosinophilic esophagitis) were excluded, unless they did not respond to medical therapy and fulfilled the Wolfson criteria.

Patients could be included even if they had underlying medical conditions, as long as they were unresponsive to medical or nutritional interventions. The inclusion criteria for this group were onset of symptoms before 2 years of age, a history of persistent food aversion or poor oral intake lasting >1 month, and abnormal feeding patterns at any time during a period of refusal. The exclusion criteria were age >6 years, parents who failed to show up for more than the diagnostic visit, pervasive developmental disorder, children who were currently tube fed, or patients presenting with caregivers other than a parent to the diagnostic or first therapeutic visit. All of the parents were treated by 1 of 4 physicians participating in a failure-to-thrive clinic at the Wolfson Medical Center, and all had exposure to the role reversal method and its principles.

After the first visit, all of the parents were invited for a first follow-up visit within 2 to 4 weeks, and the visits after this time point were individualized (every 2–4 weeks based on progress). Each visit lasted 30 to 45 minutes, and the parents met with the treating physician and dietitian simultaneously in each session in the same room. At the time of inclusion, the parents filled a nonvalidated questionnaire that comprised 6 categories of questions related to demographic and feeding behaviors (yes/no questions regarding presence or absence of food refusal, presence or absence of anticipatory gagging, presence or absence of prandial or postprandial vomiting, use of forced feeding, and distraction). Additional semiqualitative questions were regarding the will and interest to eat specific types of food deliveries—bottle, spoon-feeding, and certain textures. Parents were also questioned regarding their perception of their child's feeding behaviors (are parents worried about their child's feeding behavior or growth rate, does the child eat better with other caregivers, etc). The questionnaires also consisted of qualitative scales that rated the attitudes of the parents toward their child's feeding disorder, and a set of multiple choice questions that recorded the frequency of feeding difficulties (always exist/sometimes/seldom). As per the method of Levy et al (11), pathological feeding was categorized as follows: nocturnal feeding, forced feeding, persecution, conditional distraction, and mechanistic feeding.

The parental questionnaire was then compared with the patient's medical record to verify the presence or absence of pathological feeding and symptoms. At the last follow-up the mother was asked to fill the same form again, and the results were cross-checked with the patient's file. This measure was believed to be necessary as we have detected underreporting of abnormal feeding behavior in questionnaires and of infant feeding difficulties, which was evaluated more thoroughly in the patient's file at each visit. Failure to thrive at baseline was defined as a decrease of 2 standard deviations or more in CDC weight-for-age curves at the time of presentation.

Back to Top | Article Outline

Treatment by Role Reversal

To apply this method, it is critical to identify all the pathological feeding behaviors used by parents. In brief, the treatment was administered by 1 of 4 physicians along with 2 dietitians, without a routine evaluation or involvement of a psychologist. The psychologist was involved in the treatment of patients if there was a global or specific parental malfunction, resistance to the method, or if the parents requested to meet with one. Food-play sessions were available once a week for 20 to 40 minutes on request of the parent or physician. In these sessions young children were seated around picnic tables with finger food, along with a nurse, with the parents observing but not intervening.

A more detailed explanation of the treatment method is presented in the online-only appendix ( In brief, therapy for all of the patients’ caregivers was performed on an ambulatory basis, and was based on modification of parental behavior and not by direct infant behavioral modification. Once the diagnosis was entertained, we explained to parents that the child's feeding behavior was a response to parental feeding behavior, which led to a vicious cycle of refusal and circumvention of refusal by abnormal parental feeding and reinforcement of refusal with increased intrusive feeding behavior. We identified the particular feeding behaviors and asked the parents to avoid all pathological feeding habits, to respect refusal, avoid placing emphasis on weight or weighing the child, and avoid reinforcing existing feeding behavior by awards punishment or explanations.

We defined full response to treatment as a response that resulted in clear modification of all 3 outcomes: normative infant feeding at end of follow-up (defined below), improvement in parental feeding patterns, and improvement or stabilization of growth curves, as assessed by the team in clinic and by responses in the follow-up questionnaires.

Full response of infant or toddler feeding required initiation of feeding by the infant or child, disappearance of refusal, increased variety of foods, and cessation of prandial or postprandial vomiting. Parental response was defined as cessation of abnormal feeding patterns. The growth response was previously defined.

The endpoint for the study was full response by 6 months, which is described in the previous section. For patients with full response by 3 months and discharged from follow-up, the last visit was used for analysis. The secondary end point was partial response, which was defined as improvement in two-third categories; 1 of the categories had to be infantile feeding.

Back to Top | Article Outline


The analysis of data was carried out by using SPSS 21.0 (IBM SPSS, Armonk, NY). Normality of distribution of continuous variables was assessed using the Kolmogorov test (cutoff at P = 0.01). All continuous variables had approximately normal distributions and were compared by response to treatment using 1-way analysis of variance, followed post hoc by Bonferroni test for pairwise comparisons. Continuous variables were compared by dichotomous variables such as the presence or absence of specific feeding pathologies using the t test for independent samples. Categorial variables were described as frequency counts presented as n (%). Categorial variables were compared by treatment response using the χ2 test. Change in specific feeding pathologies before versus after treatment was compared using McNemar test. All tests are 2-sided and considered significant at P < 0.05.

Back to Top | Article Outline


We enrolled 39 patients of whom 7 were excluded, 6 because of lack of follow-up after the first visit and 1 because the final diagnosis was vulnerable child syndrome and not an IFD. The median age was 16.5 months (range 14–69). Demographic data are presented in Table 3. The most common trigger was transition (from bottle to spoon, breast-feeding to bottle or spoon), followed by mechanistic feeding (ie, clockwork feeding at specific intervals irrespective of child's will to take feed). Pathological feeding was present in all of the patients at onset; the most commonly used was mechanistic feeding, followed by persecution and forced feeding. Active food refusal was documented in approximately 90% of patients and the rest had poor feeding without clear refusal.



There was a specific improvement in parental feeding in all categories (mechanistic 10/27, nocturnal 12/12, forced feeding 13/14, persecution 11/21, and distraction 14/30; Fig. 1). Parental perception of inadequate feeding was present in 88% of participants at initiation of treatment and was still present in 41% at follow-up (P = 0.002). At follow-up, 28% of parents still practiced abnormal feeding.



A response to therapy was seen in 78% of patients. A full response as defined was obtained in 53% and a partial response in 25% of the patients. No specific demographic or triggering factors were associated with partial or full response (including age at onset, age at therapy), although we did notice differences in the full success rates between therapists (data not shown).

Back to Top | Article Outline


The present study is unique because for the first time we have published details about the principles and techniques we used in the treatment of IFDs, based on the behavioral role reversal method we have developed.

In the present prospective study, role reversal resulted in improvement in 78% of children diagnosed as having IFD. Although the overall response rate was similar to our previous study, the full response rate was lower (53% vs 73%.) than that in our previous study. This was manifested by unresolved residual abnormal feeding patterns among some of the parents. We also noticed prominent variability between participating physicians in the success rate, which may reflect less emphasis on cessation of pathological feeding early in the treatment. All forms of pathological feeding had significantly improved in response to treatment; the complete cessation of nocturnal feeding and mechanistic feeding was not just the result of the treatment, because the ability to use maladaptive feeding behaviors such as nocturnal feeding is technically more difficult as the infant grows older. Specifically, parents responded most commonly by stopping distraction during feeding and by stopping forced feeding, whereas persecutory feeding was harder to change, as evidenced by residual persecutory behaviors even among some parents with marked improvement in infant feeding. The ability to “let go” and avoid dealing with a child's food intake or feeding is also time dependent; as the child improves with feeding, the parental anxiety seems to decline. We did not identify any factor that was predictive of failure, but the cohort is probably extremely small to detect predictors of failure. In our previous article, performed in a larger cohort (10), age >3 years at first treatment was predictive of less likelihood to achieve full response.

Several different methods for the treatment of IFD have been developed and used; among these are extinction therapy, contingency management, and modeling. In brief, extinction involves repeated exposure to the rejected item until acceptance is reached. Contingency management involves positive or negative rewarding according to the child's will to consume undesired food. Modeling is a process of reinforcing the child to eat by imitating the caregiver (15–18). These methods can be traumatic (extinction therapy) or reinforce negative behaviors by secondary gain (contingency management), and all do not attempt to treat the underlying causes that led to the development of IFDs. The treatment of IFDs to date has been problematic, and has lacked uniform reproducible structure. Role reversal therapy (previously termed avoidance transfer) is a behavioral method that can be adapted by medical personnel such as pediatricians, gastroenterologists, or dietitians. Its goal is to reestablish normative age-appropriate feeding by allowing the parents to relinquish the drive and responsibility to feed the child. The infant or child develops the drive and autonomy to eat and decides when to feed and what to eat (reversal of roles in feeding). This method aims to restore confidence in feeding and develop a normal hunger satiety gratification cycle, which is 1 of the prime drivers for feeding behavior in normal children.

In conclusion, we have shown that this method can be used with success in the ambulatory setting by physicians and dietitians. This study reflects real-life results and demonstrates that infantile feeding can be improved by decreasing intrusive feeding.

Back to Top | Article Outline


1. Chatoor I, Ganiban J. Food refusal by infants and young children: diagnosis and treatment. Cogn Behav Pract 2003; 10:138–146.
2. Wright C, Parkinson K, Drewett R. How does maternal and child feeding behavior relate to weight gain and failure to thrive? Data from a prospective birth cohort. Pediatrics 2006; 117:1262–1269.
3. Chatoor I, Surles J, Ganiban J, et al. Failure to thrive and cognitive development in toddlers with infantile anorexia. Pediatrics 2004; 113:440–447.
4. Ramsay M, Gisel E, Boutry M. Non organic failure to thrive: growth failure secondary to feeding skills disorder. Dev Med Child Neurol 1993; 35:285–297.
5. Mathisen B, Worrall L, Masel J, et al. Feeding problems in infants with gastro-esophageal reflux disease: a controlled study. J Pediatr Child Health 1999; 35:163–169.
6. Heyman PE. Gastroesophageal reflux: one reason why baby won’t eat. J Pediatr 1994; 125:S103–S109.
7. Hill D, Heine R, Cameron D, et al. The natural history of intolerance to soy and extensively hydrolyzed formula in infants with multiple food protein intolerance. J Pediatr 1999; 135:118–121.
8. Staiano A. Food refusal in toddlers with chronic diseases. J Pediatr Gastroenterol Nutr 2003; 37:225–227.
9. Reilly SM, Skuse DH, Wolke D, et al. Oral motor dysfunction of children who fail to thrive: organic or non organic? Dev Med Child Neurol 1999; 41:115–122.
10. Levine A, Bachar L, Tsangen Z, et al. Screening criteria for diagnosis of infantile feeding disorders as a cause of poor feeding or food refusal. J Pediatr Gastroenterol Nutr 2011; 52:563–568.
11. Levy Y, Levy A, Zangen Z, et al. Diagnostic clues for identification of nonorganic vs organic causes of food refusal and poor feeding. J Pediatr Gastroenterol Nutr 2009; 48:355–362.
12. American Psychiatric AssociationDiagnosis and Statistical Manual of Mental Disorders. 4th ed.Washington, DC:American Psychiatric Association; 2000.
13. Chatoor I. Feeding disorders in infants and toddlers: diagnosis and treatment. Child Adolesc Psychiatr Clin North Am 2002; 2:163–183.
14. Diagnostic Classification of Mental Health and Development Disorders of Infancy and Early Childhood. Revised ed.Washington, DC:Zero to Three; 2004.
15. Ahearn W. Effects of two methods of introducing foods during feeding treatment on acceptance of previously rejected items. Behav Inerv 2002; 17:111–127.
16. Gentry J, Luiselli J. Treating a childs selective eating through parent implemented feeding intervention in the home setting. J Dev Phys Disab 2008; 20:63–73.
17. Binnendyk L, Lucyshyn J. A family centered positive behavioral support approach to the amelioration of food refusal behavior. J Posit Behav Interv 2009; 11:47–62.
18. Nock M. A multiple baseline evaluation of the treatment of food phobia in a young boy. J Behav Ther Exp Psychiatry 2002; 33:217–225.

food refusal; infantile feeding disorder; Wolfson criteria

Supplemental Digital Content

Back to Top | Article Outline
© 2014 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology,