Evidence to support treatment options for children with swallowing and feeding disorders: A systematic review

Abstract

PURPOSE:

A rise in pediatric patients with swallowing and feeding problems has resulted in increased interest in multidisciplinary treatments to address these issues. This evidence based systematic review (EBSR) examined the published evidence for the use of common strategies used by clinicians across disciplines to treat pediatric swallowing and feeding problems.

METHODS:

A systematic search of 10 electronic databases was completed to identify relevant, peer reviewed literature published in English prior to December 2015 reporting original data that addressed at least one of the five identified clinical questions.

RESULTS:

Sixty-one studies of varying methodological quality were included. The majority of the included studies (60/61) focused on the use of behavioral therapies to remediate swallowing and feeding disorders in children and reported mixed findings across all of the targeted outcomes.

CONCLUSION:

There is insufficient quantity of evidence to determine the effects of oral motor, sensory, and pharmaceutical therapies on functional feeding outcomes in pediatric populations. A larger body of phase 1 evidence is available that establishes the efficacy of behavioral strategies to treat some swallowing and feeding difficulties in small cohort and single subject studies. This analysis identified limited high quality (phase 4) research articles that establish the efficacy and benefit of joint nutrition and behavior intervention programs and systematic desensitization and operant conditioning behavioral therapy approaches to improve functional feeding and swallowing outcomes in children.

1. Introduction

Pediatric feeding and swallowing disorders have the potential for significant morbidity and even mortality due to multiple body system involvement [1]. Successful remediation of pediatric feeding and swallowing disorders requires a clear understanding of the contributing factors from accurate diagnosis by a multi-disciplinary team of professionals that may include dieticians, nurses, physicians (developmental pediatrician, gastroenterologist, pulmonologist, and/or otolaryngologist), therapists (speech-language pathologists, occupational therapists, board certified behavior analysts (BCBA), and/or physical therapists), psychologists, social workers, and caregivers [2, 3, 4]. The multidisciplinary feeding team contributes to the number of treatment modalities (oral motor, sensory, behavioral, and pharmaceutical) recommended to address the multisystem nature of pediatric feeding and swallowing disorders [1, 5, 6, 7].

Pediatric swallowing disorders (dysphagia) and feeding disorders have distinctive diagnostic criteria. Pediatric dysphagia is a skill-based disorder that interrupts the typical, coordinated actions of the swallowing mechanism. Pediatric feeding disorders are best described with the revised classification of avoidant/restrictive food intake disorder (ARFID) [8]. ARFID describes children that may or may not be a normal weight, who have restricted eating patterns that are not related to body image and whose lives and the lives of their families are significantly impacted by their restricted oral intake [9].

Multidisciplinary swallowing and feeding teams often prescribe a number of different treatment modalities (oral motor, behavioral, sensory, and pharmaceutical) to address pediatric feeding and swallowing problems, once the overriding medical issues (when present) have been ameliorated. There are also treatment programs available that utilize two or more of the listed treatment modalities [10, 11]. While various treatment modalities are utilized in feeding clinics, the evidence that demonstrate their efficacy and support their use for various populations are not readily known. For these reasons, the authors felt this study was necessary to document the levels of evidence available to support common treatment modalities used to address feeding and swallowing problems in medically stable patients.

There have been previous evidence based systematic reviews published that evaluated some of the treatment modalities in isolation [5, 12, 13, 14, 15]. Previous reviews of oral motor interventions on swallowing in preterm infants [12] and children [13] found that for preterm populations some oral motor interventions have preliminary evidence to show limited benefit for enhancing feeding/swallowing skills [12] and for children there is insufficient evidence to determine the effect of oral motor interventions on remediation of swallowing problems [13]. In both reviews on oral motor interventions, there was specific discussion regarding the methodological limitations of the current evidence available on this modality. Angell (2010) reviewed treatment methods to address selective eating and tactile sensitivity and describes them along with the limited published evidence available to support them [14]. Finally, McGrath and colleagues (2010) primarily report on the behaviorally based treatments and the available evidence to support them for children with severe feeding issues, children who do not eat orally, and children who are dependent upon gastrostomy tube for nutritional support [15]. This study represents the first evidence based systematic review (EBSR) to cover all studies on one or more of the most commonly prescribed treatment modalities: oral motor, behavioral, sensory, and pharmaceutical.

The primary purpose of this EBSR was to examine the state of the evidence for the use of oral motor, behavioral, sensory, and/or pharmaceutical interventions to treat pediatric feeding problems. Specifically, the effects of the various treatment modalities on the functional feeding outcomes of feeding behavior(s), volume of oral intake, variety of oral intake, days to wean tube feeding dependence, weight gain, and/or growth in pediatric populations aged six months to eighteen years were evaluated. Additionally, the goal was to draw attention to the possible gaps in the current evidence base. In developing the clinical questions, the impact of each treatment modality applied individually was considered. Thus, the first four clinical questions of this EBSR were:

What is the effect of oral motor interventions on functional oral feeding outcomes in children? Oral motor interventions included exercises and activities that were designed to influence the action(s) of the tongue, lips, soft palate, jaws, larynx, and/or respiratory muscles for improved strength, tone, range of motion, or coordination during feeding/swallowing and included traditional muscle exercises (active or passive), stretching, and/or sensory stimulation to the articulators and related structures [13].

What is the effect of sensory-based interventions (exercises and activities that are designed to promote sensory integration such as using a brushing or vibration protocol, a weighted vest, or deep pressure in the joints) on functional oral feeding outcomes in children [16]?

Figure 1.

Methodology flow chart.

What is the effect of behavioral-based interventions (including approaches that incorporate techniques with elements from operant condition- ing, systematic desensitization, etc.…) on func- tional oral feeding outcomes in children? [17]

What is the effect of pharmaceutical-based interventions (including any pharmaceutical agent prescribed by a physician for the documented purpose of improving oral intake) on func-tional oral feeding outcomes in children?

Due to the known benefits of treating pediatric feeding and swallowing problems within a multidisciplinary team framework [18], the search criteria was expanded in the final question to include all studies that utilized two or more of the targeted modalities (oral motor, behavioral, sensory, and/or pharmaceutical interventions).

What is the effect of applying mixed modality interventions on functional oral feeding outcomes in children?

2. Methods

A systematic search of the literature was undertaken by a multidisciplinary team of therapists with extensive experience treating feeding and swallowing disorders in pediatric patients, two speech-language pathology graduate students, and one speech-language pathology undergraduate student in accordance with methodology that was previously described [19]. All students involved in this project were under the direct supervision of the first author. Ten different databases (PubMed, MEDLINE (Ovid), CINAHL (Cumulative Index to Nursing and Allied Health), PsycINFO, ERIC (Education Resource Information Center), Science Citation Index (Web of Science), Social Science Citation Index (Web of Science), Academic OneFile, Health Source: Nursing/Academic, & PsycArticles (NAAL)) were systematically searched between March and August 2014 and then again in November 2015. Search terms included systematic combinations of the following terms in order of population, disorder, and intervention: (population terms) pediatrics, infants, toddlers, children, adolescents, (disorder terms) dysphagia, swallowing disorder, feeding disorder/problem, volume limiting, limited variety, (intervention) oral-motor therapy, sensory-based therapy, behavior-based therapy, pharmaceutical therapy, mixed therapeutic interventions, and their related Medical Subject Headings (MeSH) terms. The same search terms & systematic combinations of those search terms were used for each database. Search terms were determined based on their relevance to the established clinical questions, identified from review of pertinent articles, and incorporated into the overall search strategy. To qualify for inclusion, a study had to be written in English, published (or in press) in a peer-reviewed journal between January 1960 and December 2015, provide original data addressing at least one of the identified clinical questions, examine infants and children from six months of age to 18 years of age with dysphagia, feeding delay, or feeding disorder as either a primary or secondary diagnosis, and feature an experimental, quasi-experimental, or single subject design. The date range of January 1960 to December 2015 was chosen in order to capture the most recently published evidence (within the last 50 years) relating to the clinical questions. Studies were excluded if they presented only information from literature review, if the age of the participants was outside of the specified range, or if the study reported on oral feeding outcomes in children without a diagnosis of either swallowing and/or feeding disorder. Inclusion and exclusion criteria were established in order to identify all of the empirical literature that addressed one or more of the established clinical questions.

Three independent reviewers identified 61 citations for final inclusion (Fig. 1). Study eligibility agreement between reviewers was accomplished through initial individual review of abstracts with follow up discussion and consensus before final inclusion. The number of initial disagreements was not documented and therefor reliability between the reviewers could not be calculated.

Using a structured rating system [20] three independent reviewers assessed all of the included articles to identify areas of methodological weakness and possible bias (Table 1), in addition to summarizing key elements (participant characteristics, treatment modality, treatment intensity and duration, and outcomes) from each study. Table 2 provides a summary of the methodological quality for each article included in this EBSR.

Final scoring agreement for assessment of methodological features was 100% as all initial discrepancies were resolved through discussion and consensus and again the number of initial disagreements was not documented and therefor reliability between the reviewers before reaching consensus could not be calculated. Tables 3–5 summarize the quality indicators of the included studies, the participant characteristics, key study variables and findings. All studies, regardless of methodological rigor were included in this analysis as a stated purpose of this EBSR was to establish the quality of evidence available to support various treatment modalities for improving oral feeding measures in children with swallowing and/or feeding disorders.

3. Results

Results are presented in order of the clinical questions posed. With regards to Clinical Question 3 (Behavioral), results are presented with subheadings in order of behavioral techniques utilized in the articles: positive reinforcement and extinction; stimulus shaping and fading; caregiver involvement and education; operant conditioning and systematic desensitization. For Clinical Question 5 (Mixed Modality), results are presented with subheadings in order of combinations utilized in the articles: behavioral and oral motor; behavioral and sensory; behavioral and pharmaceutical; behavioral, oral motor, and sensory.

3.1 Clinical Question 1 (Oral Motor)

One of the 61 included studies examined clinical question one [21]. The 20 participants in this experimental, randomized control trial study had profound intellectual disability or neuromuscular disorders, lived in a residential care facility, and were identified by the residential staff as being problem feeders. The authors implemented an oral motor program that included manual intervention, direct guidance, stretches, pressure, and oral desensitization administered 30–40 minutes daily, five times a week for nine weeks for the experimental group ( $n=$ 10). The control group continued with their previously prescribed therapeutic plan of care. Only nine subjects from the experimental group and two from the control group had both pre and post treatment evaluations of oral motor functioning. Analysis of the nine experimental subjects that had pre and post treatment oral motor function testing revealed no significant change following treatment. Statistical analysis of the full sample ( $n=$ 20) revealed no significant difference in body weight gains between the experimental and the control group at discharge [21]. See Table 3 for details.

3.2 Clinical Question 2 (Sensory)

None of the 61 articles included examined only sensory-based interventions.

Table 1

Quality indicators

Research design	Research protocol	Blinding	Type of sample	For RCTs only: Analysis by intention to treat	Evidence of treatment fidelity	$p$ value	Precision
*Randomized controlled trial Cohort Study	*Description of research protocol sufficient for replication	*Yes, assessors were blinded	*Sample was randomized into treatment conditions with sufficient description for replication	*Yes, analysis by intention to treat	*Yes, authors provide evidence of treatment fidelity	*Significance reported w/ $p$ values (or provided sufficient information to calculate)	*Effect size AND confidence intervals were reported or there was sufficient information to calculate
Retrospective case control or single subject design	Description of research protocol insufficient to allow for replication or not explicitly stated	No, assessors were not blinded or it was not explicitly stated in article	Sample was randomized into treatment conditions but there was not sufficient description for replication	No, analysis not conducted by intention to treat	No, authors do not provide evidence of treatment fidelity	Significance values not reported (no $p$ values) and there was not sufficient information provided to calculate	Effect size OR confidence interval reported/calculable
Case series Case Study			Convenience sample, handpicked sample, or allocation of subjects was not described				No information on effect size or confidence interval was reported/calculable

Adapted from Gosa, Schooling & Coleman [19]; RCT $=$ Randomized controlled trial; * indicates highest level of quality in each category.

Table 2

Summary of quality indicators

Citation	Total quality	Design	Intent to	Protocol	Blinding	Sample	Fidelity	Significance	Precision
	indicators		treat
Clinical Question 1: Oral Motor
Ottenbacher, Scoggins & Wayland (1981) [21]	3 of 8	$+$	$-$	$-$	$+$	$-$	$+$	$-$	$-$
Clinical Question 3: Behavioral
Blisset and Harris (2002) [27]	0 of 8	$-$	$-$	$-$	$-$	$-$	$-$	$-$	$-$
de Moor, Didden & Korzilius (2007) [28]	0 of 8	$-$	$-$	$-$	$-$	$-$	$-$	$-$	$-$
Luiselli and Luisellli (1995) [37]	0 of 8	$-$	$-$	$-$	$-$	$-$	$-$	$-$	$-$
Pizzo, Williams, Paul & Riegel (2009) [44]	0 of 8	$-$	$-$	$-$	$-$	$-$	$-$	$-$	$-$
Bachmeyer et al. (2009) [23]	1 of 8	$-$	$-$	$-$	$-$	$+$	$-$	$-$	$-$
Bandstra, Phillips, and Flowerdew (2011) [24]	1 of 8	$-$	$-$	$-$	$-$	$-$	$+$	$-$	$-$
Dovey and Martin (2012) [29]	1 of 8	$-$	$-$	$-$	$-$	$-$	$-$	$+$	$-$
Gale, Eikeseth, and Rudrud (2010) [30]	1 of 8	$-$	$-$	$-$	$-$	$-$	$+$	$-$	$-$
Greer, Dorow, Williams, and McCorkle (1991) [31]	1 of 8	$-$	$-$	$-$	$-$	$-$	$+$	$-$	$-$
Luiselli (2000) [36]	1 of 8	$-$	$-$	$-$	$-$	$-$	$+$	$-$	$-$
Rivas, Piazza, Patel, and Bachmeyer (2010) [46]	1 of 8	$-$	$-$	$-$	$-$	$-$	$+$	$-$	$-$
Schadler, Suss-Burghart, Toschke, von Voss, and von Kries (2007) [48]	1 of 8	$-$	$-$	$-$	$-$	$-$	$-$	$+$	$-$
Seiverling, Williams, Sturmey, and Hart (2012) [49]	1 of 8	$-$	$-$	$-$	$-$	$-$	$+$	$-$	$-$
Vaz, Volkert, and Piazza (2011) [55]	1 of 8	$-$	$-$	$-$	$-$	$-$	$+$	$-$	$-$
Binnedyk and Lucyshyn (2008) [26]	2 of 8	$-$	$-$	$+$	$-$	$-$	$+$	$-$	$-$
Groff, Piazza, Zelleny, and Dempsey (2011) [33]	2 of 8	$-$	$-$	$+$	$-$	$-$	$+$	$-$	$-$
Groff, Piazza, Volkert, and Jostad (2014) [32]	2 of 8	$-$	$-$	$+$	$-$	$-$	$+$	$-$	$-$
Johnson, Foldes, DeMand, and Brooks (2015) [34]	2 of 8	$-$	$-$	$-$	$-$	$-$	$+$	$+$	$-$
LaRue et al. (2011) [35]	2 of 8	$-$	$-$	$+$	$-$	$-$	$+$	$-$	$-$
Najdowski, Wallace, Doney, and Ghezzi (2003) [40]	2 of 8	$-$	$-$	$+$	$-$	$-$	$+$	$-$	$-$
Opipari-Arrigan, Powers, Quittner, and Stark (2010) [41]	2 of 8	$-$	$-$	$+$	$-$	$-$	$-$	$+$	$-$
Patel, Piaza, Martinez, Volkert, and Santana (2002) [42]	2 of 8	$-$	$-$	$+$	$-$	$-$	$+$	$-$	$-$
Penrod & VanDalen (2010) [43]	2 of 8	$-$	$-$	$+$	$-$	$-$	$+$	$-$	$-$
Rivas, Piazza, Roane, Volkert, et. al. (2014) [47]	2 of 8	$-$	$-$	$+$	$-$	$-$	$+$	$-$	$-$
Sharp, Jaquess, Morton, and Miles (2011) [51]	2 of 8	$-$	$-$	$-$	$-$	$-$	$+$	$+$	$-$
Stark et al. (1993) [52]	2 of 8	$-$	$-$	$-$	$-$	$-$	$+$	$+$	$-$
Tarbox, Schiff, and Najdowski (2010) [54]	2 of 8	$-$	$-$	$+$	$-$	$-$	$+$	$-$	$-$
Volkert, Peterson, Zeleny & Piazza (2014) [56]	2 of 8	$-$	$-$	$+$	$-$	$-$	$+$	$-$	$-$
Wilkins, Piazza, Groff, and Vaz (2011) [57]	2 of 8	$-$	$-$	$+$	$-$	$-$	$+$	$-$	$-$
Wilkins, Piazza, Groff, Volkert, Kozisek & Milnes (2014) [58]	2 of 8	$-$	$-$	$+$	$-$	$-$	$+$	$-$	$-$
Sharp, Burrell, and Jacquess (2013) [50]	3 of 8	$+$	$-$	$-$	$-$	$+$	$-$	$+$	$-$
Stark et al. (1996) [53]	3 of 8	$-$	$-$	$+$	$-$	$-$	$+$	$+$	$-$
Powers et al. (2005) [45]	5 of 8	$+$	$+$	$-$	$-$	$+$	$-$	$+$	$+$
Benoit, Wang, and Zlotkin (2000) [25]	6 of 8	$+$	$+$	$-$	$+$	$+$	$+$	$+$	$-$
Marshall, Hill, Ware, Ziviani, and Dodrill (2015) [38]	7 of 8	$+$	$+$	$+$	$-$	$+$	$+$	$+$	$+$
Morawska, Adamson, Hinchcliffe, and Adams (2013) [39]	7 of 8	$+$	$+$	$+$	$-$	$+$	$+$	$+$	$+$

Table 3

Clinical question 1 (oral motor) data summary

Citation	N	Age in months (Mean)	Sex	Diagnoses as reported in article	Intervention	Treatment frequency and duration	Outcome Measure(s)
Ottenbacher, Scoggins & Wayland (1981) [21]	20	5.0–21.6 (11.5)	12 M 8 F	Intellectual disability Neuromuscular disorders Cerebral palsy	Oral motor: Inhibition of abnormal oral and postural reflexes, facilitation of normal muscle tone, and the desensitization of the oral region	30 to 40 minutes/day 5 days/week 9 weeks	Weight Gain

$N=$ number of subjects; M $=$ male; F $=$ female.

Table 4

Clinical question 3 (behavioral) data summary

Citation	N	Age in months (Mean)	Sex	Diagnoses	Intervention	Treatment frequency and duration	Outcome Measure(s)
Adamson, Morawska, and Sanders (2013) [22]	96	18–72 (M $=$ 31)	45 M 51 F	Feeding problems	Behavioral: Hassle free Mealtimes Triple P	4, two-hour group sessions 3 telephone sessions 1 group session Provided over 8 weeks	Child feeding behaviors (problematic and general) Mealtime parenting strategies Parental cognition Mealtime diary
Bachmeyer et al. (2009) [23]	4	42–60 (M $=$ 48)	2 M 2 F	Feeding problems Hernia repair times 2 DD GERD G-tube dependence OMD BPD Hypotonia Left sided facial nerve paralysis Visual impairment	Behavioral: Escape and attention extinction Reinforcement	30–45 minute meals 5–12 meals/day with 3–12 sessions per meal during 2 hour outpatient visits Duration: NR	Acceptance Appropriate mealtime behaviors
Bandstra, Phillips, and Flowerdew (2011) [24]	490	12–164 (40.5)	285 M 205 F	Behavioral feeding issues Cardiovascular or respiratory complications Renal concerns Neurological concerns Cystic fibrosis GI concerns Immune disorders Poor growth	Behavioral: Not otherwise specified	Frequency: NR Initial assessment and follow up 1–2 years after	Number of physician visits Behavioral treatment effects Associated costs
Benoit, Wang, and Zlotkin (2000) [25]	64	4–36 (M $=$ 20)	37 M 27 F	G-tube dependency	Behavioral Structured schedules and routines (both groups) Extinction (only one group)	1 weekly treatment session 7 weeks 4 follow-up visits	Problem feeding behaviors Time to wean feeding tube dependence Mouth clean Weight Length
Binnedyk and Lucyshyn (2008) [26]	1	72	M	ASD Food selectivity	Behavioral: Positive behavior support Daily eating schedule General case programming Stimulus fading Positive contingency statement Prompting with prompt fading Contingent praise Contingent access to preferred toy or activity Escape extinction De-escalation procedure	Treatment 2–4 times/week Parent training 1–2 times/ week 15 sessions Administered over 14 weeks	Volume of intake with therapist Volume of intake with parent Generalization Self-initiation

Table 5

Clinical question 5 (mixed modality) data summary

Citation	N	Age in months (Mean)	Sex	Diagnoses	Intervention	Treatment frequency and duration	Outcome Measure(s)
Addison et al. (2012) [59]	2	12–36 (M $=$ 24)	1 M 1 F	GERD Hypertension Asthma CHD FTT DD Dysphagia Mitochondrial disorder	Behavioral Sensory	Meals 5x/day for 30–45 minutes Duration NR	Acceptance Inappropriate behaviors Grams consumed
Berger-Gross, Colletti, Hirschkorn, Terranova, and Simpser (2004) [60]	3	84–114 (M $=$ 96)	1 M 2 F	FTT GERD Depression Anxiety FTT Probably Myelofibrosis G-tube dependency Esophageal dysmotility TEF	Behavioral Pharmaceutical	Inpatient feeding program for 20–26 weeks	Calorie intake Weight gain
Clawson, Kuchinski, and Bach (2007) [62]	8	19–55 (M $=$ 32)	4 M 4 F	Oral dysphagia H/o prematurity FTT	Behavioral Oral Motor	6 hours/day with 4 therapeutic meals/day 5x/week Average length of stay 29 days Follow up assessments at 1, 4, 7, and 12 months post treatment	Acceptance Mouth clean Inappropriate behaviors Volume consumed Rate of consumption Weight gain Caregiver skills
Clawson, Palinsky, and Elliot (2006) [61]	3	15–42 (M $=$ 36)	1 M 2 F	Goldenhar syndrome	Behavioral Oral Motor Sensory	4 meals/day 5 days/week Duration average of 37 days (range 28–48 days)	Anthropometric data Acceptance Length of meal Grams consumed Percentage of bites expelled Inappropriate meal time behaviors Time to wean from tube feeding
Davis, Bruce, Cocjin, Mousa, and Hyman (2010) [18]	9	7–48 M $=$ 27.5	NR	Dysphagia GERD Ring 21 abnormality VSD CP Trisomy 21 G tube dependency	Behavioral Pharmaceutical	3, 1-hour treatment sessions 14 weeks total time in program	Weight gain Calorie intake Time to wean from tube feedings
Douglas & Harris (2001) [63]	73	NR	NR	FTT Food selectivity Feeding delays	Behavioral Sensory	12 total sessions administered biweekly over 24 weeks	Reduction in tube feedings Weight percentile Volume of food and drink consumed Selectivity Behaviors towards foods

3.3 Clinical Question 3 (Behavioral)

Thirty-seven of the 61 included studies examined clinical question three [22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58]. Of 919 participants across 37 studies, a majority of participants (86%) were between two and seven years old [22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58]. Overall sample sizes were small (1–13 participants) [23, 26, 27, 28, 30, 31, 32, 33, 35, 36, 37, 40, 42, 43, 44, 45, 46, 47, 49, 51, 52, 53, 54, 55, 56, 57, 58], however some studies had larger sample sizes ranging from 24 to 490 participants [22, 24, 25, 29, 34, 38, 39, 41, 48, 50]. The majority of the reports examined the prescribed intervention’s impact on volume and variety of oral intake as well as weight gain. The studies varied greatly in reported frequency and duration of treatment. One study, the largest study identified by this systematic review, did not specify the types of behavioral interventions used in their protocol and instead assessed the impact of unspecified behavioral therapy techniques on outcome success and cost in a retrospective chart review of 490 participants [24]. The results are presented with subheadings in order of behavioral techniques utilized in the articles: positive reinforcement and extinction, stimulus shaping and fading, caregiver involvement and education, and operant conditioning and systematic desensitization.

3.4 Positive reinforcement and extinction

Eleven of the 37 studies reported on the use of escape extinction (behavioral technique used to treat behaviors that are maintained by escaping or avoiding; removing the possibility of escaping from the challenging stimulus) and reinforcement (the use of any stimulus which should increase the probability of a desired response, can be positive or negative) in single-subject and small cohort designs and their impact on acceptance, intake, and presence of inappropriate behaviors at mealtime [23, 26, 27, 28, 30, 31, 32, 33, 36, 40]. Patel and colleagues reported that differential reinforcement (advantages positive reinforcement to distinguish appropriate behavior from inappropriate behavior; used to increase appropriate behavior and decrease inappropriate behavior) alone was not effective for increased oral intake, but escape extinction in combination with differential reinforcement did increase oral intake and mouth clean (upon inspection, no visible food or liquid in the subject’s mouth following intake without expulsion) in a small sample ( $n=$ 3) [42]. Larue et al. similarly found that using escape reinforcement (negative reinforcement technique, utilizing escape) alone did not increase mouth clean or decrease inappropriate behaviors; change in these behaviors was dependent on the application of escape extinction regardless of the presence or absence of reinforcement ( $n=$ 5) [35]. Penrod and VanDalen corroborated the need for escape prevention in the reduction of problematic hyper food selectivity and reported that increased acceptance of non-preferred foods depended on escape prevention in the form of non-removal of the spoon plus re-presentation of expelled bites ( $n=$ 3) [43]. Bachmeyer et al. found participants decreased inappropriate mealtime behaviors and increased acceptance of food utilizing an escape and attention extinction (removal of ability to escape and removal of attention) combination ( $n=$ 4) [23]. Wilkins and colleagues also found that a chin prompt (gentle pressure applied in upward motion to child’s chin and lower lip by the feeder/therapist) in combination with representation of expulsion per bite decreased rates of expulsion and increased volume consumed ( $n=$ 4) [57]. Pizzo et al. examined the effects of an intensive 4–5 day behavioral intervention that included meal and taste probing sessions with escape extinction for a small cohort ( $n=$ 3) and found there was an overall increase in the volume and variety of foods eaten and a decrease in inappropriate mealtime behaviors that was maintained at 1-month post treatment [44]. Gale et al. reported that non-contingent negative reinforcement (positive reinforcement not related to the child producing the target behavior; reinforcement delivered on a predetermined time schedule), positive reinforcement and extinction (disappearance of a learned behavior when that behavior is not reinforced) of disruptive behavior lead to a decrease in undesirable behavior during mealtime and an increase in volume of food eaten in a small cohort of toddlers with autism spectrum disorder ( $n=$ 3) [30].

Blisset and Harris described the effects a highly structured, behaviorally based intervention program that integrated positive and negative reinforcement, self-feeding, parent education, structured mealtimes, and repeated presentation of novel foods in their case study report. Improvements in variety and volume of oral intake, reduction in undesirable behaviors associated with feeding, improved parent-child interactions during mealtimes and increases in both height and weight were reported after completing the intervention and at specified follow-up points, up to three years post treatment [27]. Volkert and colleagues utilized reinforcement, escape extinction, and caregiver training in an outpatient treatment program to enhance mastication and self-feeding skills for three study participants and found mastication, mouth clean, and acceptance increased with early swallow decrease for all subjects [56]. Rivas et al. looked at self-feeding skills utilizing a similar therapeutic approach (reinforcement, escape extinction, caregiver training) to enhance self-feeding skills. They reported that although the probability of self-feeding increased for two subjects as the number of caregiver-fed bites increased, the remaining subject did not self-feed consistently, even when the ratio of self-fed to caregiver-fed bites was 1:5 [47]. Greer et al. conducted consecutive studies in which weight gain and intake were monitored in both a child who had never swallowed food or liquid secondary to dysphagia and a child who chronically refused food following application of treatment with positive reinforcement and peer modeling. These authors reported an increase in the volume and variety of accepted foods, as well as significant weight gain with the participant in study one having weaned from tube feeding by two months’ post treatment [31].

3.5 Stimulus shaping and fading

Eleven of the 37 articles reviewed examined shaping (behavior training that uses differential reinforcement of successive approximations of the desired behavior) and stimulus fading (initial prompting to perform behavior is gradually withdrawn, ensuring the child does not become dependent on the stimulus) of placement, texture, taste and quantity combined with ranging levels of reinforcement to eliminate food refusal or discontinue tube feeding [28, 29, 32, 33, 36, 37, 46, 48, 49, 51, 58]. De Moor et al. reported generalization of oral acceptance in their case series for a variety of food items during follow up at three to twelve months with tube feeding discontinued in each subject ( $n=$ 5) [28]. Luiselli and Luiselli reported clear increases in oral intake with the reduction of gastrostomy-tube feedings and addition of oral feeding meals in their case study [37]. Luiselli and Luiselli used visual cueing, demand fading and positive reinforcement with a subject who was weaned from tube feeding to full oral feeding and gained more than two pounds from pre-treatment [36]. In a retrospective analysis, Schadler et al. examined the effect of positive reinforcement, escape extinction, and flooding (rapid and excessive exposure to a feared stimulus) on oral intake with 86 ex-premature participants with feeding disorders and found that $\sim$ 60% of them presented with some improvement in oral intake [48]. Seiverling et al. found an increased acceptance of bites (50% at baseline, at least 86% at follow up in single subject design) and a decreased number of disruptive responses from children in their small study ( $n=$ 3) [49].

Groff et al. and Rivas et al. had similar findings assessing acceptance, mouth clean, and inappropriate behaviors during mealtime in single subject and small cohort designs [33, 46]. For Groff et al. acceptance was 0% during the reversal to baseline with both a spoon and cup. During fading steps: acceptance and mouth clean remained high and inappropriate behavior remained low ( $n=$ 1) [33]. For Rivas et al., fading plus escape extinction resulted in low rates of inappropriate behaviors and high rates of acceptance (Study 1: $n=$ 2/ Study 2: $n=$ 1) [46]. Groff et al., looked to increase acceptance and mouth clean and decrease inappropriate meal time behaviors with extinction, syringe fading and planned ignoring in their later case study; baseline mean for inappropriate behaviors decreased from 31 to 3.8 per minute following extinction intervention. Acceptance and mouth clean increased to 98% given trials across five meal blocks daily [32].

Wilkins et al., combined escape extinction techniques (including non-removal of a spoon), bite representation, and a Nuk brush to increase mouth clean and acceptance, and decrease packing (holding food in the mouth and cheeks) and expulsions per bite in 12 subjects with generalized feeding complaints, developmental delays and G-tube dependency. All participants transitioned from a Nuk brush to a spoon during either an intensive outpatient or day-treatment admission or during outpatient follow up. Their findings suggest that utensil manipulation during initial treatment of children with severe feeding problems may be warranted [58].

Sharp et al. found retrospectively that using a combination including stimulus fading and non-contingent access provided significant improvements regarding food variety, consumption, and appropriate mealtime behaviors in a case series report ( $n=$ 13). They found that all participants increased their overall intake by at least 12 foods (relatively distributed across all food categories) and treatment was effective in reducing mealtime stress during food presentations [51]. Dovey and Martin found that all subjects ( $n=$ 24) increased in height, weight, and BMI when they utilized a program that incorporated behavioral principles in teaching, parent education, and concurrent presentations of preferred foods with target foods [29].

3.6 Caregiver involvement and education

Thirteen of the 37 studies focused on parent education and involvement in combination with behavioral strategies such as reinforcement, escape extinction and fading [22, 25, 26, 34, 39, 40, 41, 45, 50, 52, 53, 54, 55]. Half of these reported on findings from case studies. Vaz et al. reported an increase in the intake of pureed foods and an increase in self-feeding in a case study with eventual parent training and carryover [55], while Najdowski et al. reported an increase in the variety of novel foods accepted when using these techniques mediated by parent involvement and in-home therapy [40]. Similarly, Binnendyk and Lucyshyn reported increased variety and calorie intake of previously accepted and recently rejected foods as well as a decreased reliance on Pediasure ${}^{\mathrm{TM}}$ , a liquid nutritional supplement, with implementation of a family-centered positive behavior support program in the home [26]. Tarbox et al. provided parent training regarding non-removal of the meal and preparation of meals without consideration of child selectivity. This single subject design found a significant increase in volume and variety of food consumption following the second implementation of the intervention with a secondary decrease in the length of meal times down to below 30 minutes at follow up [54].

Stark and colleagues prospectively implemented parent education provided by clinical psychologist, pediatric dietician, and physician with personalized nutritional and behavioral management strategies, including relaxation strategies and token reward system to improve feeding in three children with cystic fibrosis. Reported outcomes included meeting calorie goals at meals and average weight gain of 0.66 kg. Utilizing positive and negative reinforcement, relaxation techniques, and parent education that included nutritional education and behavior management training [52], Stark et al. reported increased participant calorie intake ( $n=$ 9) and significant weight gain as compared to a control group over six weeks of treatment with generalization reported at six month follow-up studies [53]. Powers et al. utilized a randomized controlled design to target increased caloric intake and nutritional content with nutritional counseling and child behavior management skills with the goal of meeting 120% recommended daily allowance (RDA) per day or greater. No one met the goal at baseline but following treatment 100% of the participants in the experimental, behavioral group met the goal and only 16% of the control group (randomized to a waiting list) met the goal [45].

Sharp, Burrell, and Jacquess introduced the Autism MEAL Plan to address caregiver training for families of children with ASD ( $n=$ 19) and included didactic workshops for skill acquisition which targeted eight key subjects of parent education: introduction, structuring meals and monitoring behaviors, ways to increase appropriate behavior, effective communication, ways to decrease inappropriate behavior during meals, methods of introducing foods, teaching self-feeding skills, and monitoring and maintaining progress. Following treatment, caregivers reported a significant reduction in overall stress and large effect sizes ( $d=$ 1.1) were seen on the Parenting Stress Index short form scores [50]. Johnson and colleagues utilized a larger sample of families ( $n=$ 14) to implement a highly structured intervention package which included individual (not group) parent training (modeling, role play, and giving clear instructions, mealtime plans, reinforcement, shaping and fading) with video modeling and feedback used for caregiver skills acquisition. Reported results included statistically significant decrease in disruptive mealtime behaviors across assessment time points and significantly reduced amounts of reported parent stress [34].

Four of the remaining studies that examined the effects of parent training utilized larger cohorts (combined $n=$ 271) [22, 25, 39, 41]. Adamson et al. found in their randomized control trial that parents who received intervention with their children reported significant improvements in mealtime behavior and in their parenting styles when compared to controls ( $n=$ 96). The intensity of disruptive child behaviors, general parenting styles, and parental efficacy were also significantly improved suggesting generalization of program skills outside of meal time with these changes observed to be present and reliable at a 6-month follow up [22]. Opipari-Arrigan et al. reviewed the effects of parent training on behavioral interventions versus the effects of nutritional education alone ( $n=$ 67) and found that behavioral intervention education plus nutritional counseling (Be-In-CHARGE!) resulted in significantly greater improvements in volume of oral intake and calories with more weight gain over 9 weeks versus nutritional education alone [41]. Similarly, Benoit et al. (2000) reviewed the benefits of nutritional counseling versus the benefits of behavioral technique training in the reduction of tube-feedings utilizing a randomized controlled trial ( $n=$ 64). At the third follow-up visit, 15 (47%) of 32 patients in behavioral therapy and none of the 32 subjects (0%) in nutritional therapy were no longer dependent on tube feeding, indicating behavior therapy is more efficacious in eliminating the need for tube feeding than nutritional counseling alone [25]. Morawska et al. used a controlled group discussion and in-home workbooks to educate parents on hassle free mealtimes ( $n=$ 44) and found that parents reported a significant decrease in challenging mealtime behaviors and length of meals with an increase in volume and variety of foods eaten at six months post education [39].

3.7 Comparison of operant conditioning approach with parent education to systematic desensitization approach with parent education

Marshall et al. conducted a randomized controlled trial to compare the effectiveness of operant conditioning (learning approach that utilizes consequences to control behavior; key components include positive and negative reinforcement and positive and negative punishment) and systematic desensitization (learning approach that engages the patient in some type of relaxation exercise and gradual exposure to increasing levels of an anxiety producing stimulus like food) approaches, both provided with a parent education component, to determine their impact on specific primary outcomes of dietary intake, dietary variety, and problem feeding behaviors. The authors found significant improvement in all primary outcomes ( $p<$ 0.05) with small to medium effect sizes when groups were combined and compared with pre and post treatment measures; however, they did not find significant differences in outcome measures between the two behavioral treatment conditions.

3.7.1 Clinical Question 4 (Pharmaceutical)

None of the 61 articles included examined only pharmaceutical-based interventions.

3.7.2 Clinical Question 5 (Mixed Modality)

Twenty-three of the 61 included studies examined clinical question five [18, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80]. Of the 395 participants across 23 studies, subjects were at various levels of dependence and were identified for the following reasons: poor oral intake, food selectivity, food refusal, inability to self-feed or a delay in oral-motor skills [18, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80]. A majority of participants (95%) were between the ages of 2 and 10 years old. Sample populations were identified by a variety of techniques and included children with highly variable developmental backgrounds: Down syndrome ( $N=$ 12) [64, 65, 76], autism spectrum disorder (ASD) ( $N=$ 67) [68, 69, 72, 78, 80]. Goldenhar syndrome ( $N=$ 3) [61]. Rett syndrome ( $N=$ 3) [73], cerebral palsy ( $N=$ 9) [18, 68, 75], rubella syndrome ( $N=$ 1) [70], and generalized feeding complications due to other major system impairments ( $N=$ 243) [59, 60, 62, 63, 66, 67, 71, 75, 77, 79, 80]. The majority of studies ( $n=$ 20, 87%) featured small sample sizes (less than 10 subjects) [18, 59, 60, 61, 62, 64, 65, 66, 67, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 80]; however, three of the included studies featured much larger sample sizes reporting $n=$ 46, $n=$ 73, and $n=$ 221 respectively [63, 68, 79]. As the studies reviewed the success of mixed intervention strategies, dosage of intervention (including frequency and duration of treatment), and intervention techniques vary greatly. Most studies implement intervention at least 5 days a week with multiple feedings throughout the day, either reporting total number of meals per subject or regularly timed scheduled feedings defined by predetermined criteria [59, 60, 61, 64, 65, 66, 67, 69, 70, 73, 75, 76, 77, 80]. See Table 5 for detailed summary.

3.8 Behavioral and oral-motor interventions

Twelve of the identified 23 studies featured a combination of behavioral and oral-motor interventions (combined $N=$ 30) [62, 64, 65, 66, 67, 69, 72, 73, 74, 75, 77, 80]. All of the studies identified utilized case study or single subject designs to describe the effects of previously discussed behavioral techniques (such as escape extinction, reinforcement, stimulus fading, and shaping) in combination with oral motor principles, techniques, or tools and described their combined impact on oral feeding outcomes in heterogeneous pediatric populations [62, 64, 65, 66, 67, 69, 72, 73, 74, 75, 77, 80].

Harding, Faiman, and Wright targeted oral versus gastrostomy tube (g-tube) intake ( $n=$ 3) with intensive parent coaching regarding textures, desensitization, and problem mealtime behaviors and reported g-tube feedings were significantly reduced or eliminated with an overall increase in positive parent behavior and meal time situations in 67% of patients (two of the three subjects in the study) [67]. Sharp et al. assessed the reduction of problem behaviors during meals and improved ability to maintain bolus within the oral cavity in one subject with problem feeding behavior and oral motor dysfunction due to diagnosis of cerebral palsy [74]. Data from their case study showed reduced rates of expulsions and increased mouth cleans with adaptation of presentation to accommodate for existing oral motor dysfunction in addition to escape extinction and negative reinforcement [74]. Utilizing similar approaches, Eckman et al. improved variety and texture of food eaten in 2 male participants by implementing a structured approach to teach chewing in combination with behavioral strategies [64]. Both participants in their report were chewing with at least 80% accuracy by discharge from therapy [64]. In a report from Vaz and colleagues escape extinction in the form of non-removal of the spoon and use of a preferred food or liquid chaser following food acceptance ( $n=$ 3) decreased packing and increased mouth cleans [80]. Qvarfordt et al. utilized guided eating (including physical and verbal prompts) to successfully increase oral coordination and participation in mealtimes in three subjects with Rett syndrome [73].

Clawson and colleagues reported on the effectiveness of the Beckman Oral Motor Program with escape extinction, reinforcement, and parent training on food acceptance, mouth clean, negative vocalizations, inappropriate behaviors and caregiver interactions for eight subjects [62]. At the end of treatment, improvements were reported for mouth opening at the presentation of food, timeliness of managing food in the mouth, toleration of longer meals, and acceptance of larger volumes of food per meal. Inappropriate behaviors were reduced and caregivers demonstrated increased ability to successfully feed their children with improvement in ideal weight seen in follow up sessions [62]. A case study from Sheppard utilized incidental teaching and positive reinforcement to teach oral motor concepts and improve sensory tolerances for eating with daily practice at meal and snack times over a two year period [77]. She reported success in the development of functional feeding skills with age appropriate textures [77].

Girolami et al. [66] and Gibbons et al. [65] documented use of a Nuk brush in an oral motor condition to increase tactile stimulation and escape extinction. Girolami et al. reported that presenting and re-presenting food with the oral motor tool (Nuk brush) was more effective in decreasing expulsions than using a spoon; however, success of the brush may be due to negative reinforcement [66]. In contrast, Gibbons et al. attributed the weaning of gastrostomy tube feeding and the elimination and tongue thrust in their subject to the use of the combined strategies, oral-motor and behavioral procedures [65]. Sharp and colleagues recruited three subjects that were all gastrostomy tube dependent and implemented a behavioral program with variable spoon presentations to assess its impact on reducing undesirable mealtime behaviors, mouth clean and expulsions [75]. The authors reported that the behavioral program was effective in decreasing inappropriate mealtime behaviors and increasing acceptance of bites; however, variable spoon presentations were necessary to reduce the level of food expulsion and increase mouth clean [75]. Penrod, Gardella, and Fernand found that antecedent-based strategies increased food consumption in the absence of escape extinction for the two subjects in their study with feeding disorders and ASD [72]. Levin et al. utilized redistribution, swallow facilitation, and swallow chaser techniques to increase mouth clean and decrease packing in children with ASD ( $n=$ 2) [69].

3.9 Behavioral and sensory intervention

Four of the 23 studies utilized principles and techniques from sensory integration in combination with behavioral intervention strategies [59, 63, 70, 78]. Luiselli and Gleason combined sensory reinforcement in the form of a light, rocking motion, and texture fading to overcome chronic food refusal ( $n=$ 1) [70]. Outcome measures included acceptance and nonacceptance behaviors of solid foods; at baseline, acceptance of solid foods was 0%. During treatment acceptance increased to above 80%. Acceptance decreased during month follow-ups, but never fell below 80%. As Luiselli’s et al. participant was a 4-year-old child with visual and hearing deficits, these findings support the notion of including sensory reinforcement techniques in treating food refusal particularly in children with documented sensory deficits [70].

Addison et al. combined individualized sensory integration therapy with escape extinction and noncontingent reinforcement as a way to increase acceptance and grams consumed and decrease inappropriate mealtime behaviors for two young children with poor nutritional intake and growth [59]. Results of functional analyses favored escape extinction plus noncontingent reinforcement for increasing acceptance, decreasing inappropriate behavior, and increasing amount of calories consumed [59]. Douglas and Harris reported on a large sample ( $n=$ 73) of children treated in an outpatient-feeding clinic with a broad approach that included systematic desensitization, operant conditioning, family therapy, positive reinforcement, appetite manipulation, sensory play, and dietary advice to reduce the need for tube feeding, improve behavior during mealtimes and increase weight percentiles, grams consumed, and caregiver satisfaction. The authors reported success in outcome measures based on individualized tailoring of the vast therapeutic approaches utilized in their clinic [63].

Most recently, Suarez retrospectively reviewed the use of sensory integration and systematic desensitization in combination with escape extinction, positive reinforcement, and parent training within a home therapy program ( $n=$ 4) [78]. Acceptance of non-preferred foods increased. All participants showed a positive trend in food acceptance over a relatively small number of treatment sessions (1–10) and were discharged after caregivers reported that they were able to facilitate their child taking at least 1 bite of every family food at every meal [78].

3.10 Behavioral and pharamceutical intervention

Two of the 23 studies reviewed the effects of pharmaceutical intervention in combination with behavioral therapy techniques in single subject designs [18, 60]. One of the two studies looked at the use of Risperidone in three children with feeding disorders [60]. While outcomes varied, all of the subjects ( $n=$ 3) showed improvement in oral intake with at least partial reduction in calories administered via tube feeding [60]. Similarly, Davis et al. utilized pharmaceutical treatment (Gabapentin and/or amitriptyline, megesterol) plus hunger provocation to assess weight gain and percent of oral intake ( $n=$ 9) [18]. Eight of the nine subjects maintained weight and never required additional tube feedings following treatment [18].

3.11 Behavioral, Oral Motor, & Sensory interventions from multidisciplinary teams

Five of the 23 articles reviewed intervention techniques that involved a multidisciplinary approach to improving feeding outcomes [61, 68, 71, 76, 79]. Clawson and colleagues utilized a large team of providers including occupational therapy, speech therapy, feeding technicians, dieticians, nutritionists, gastroenterologists, nursing staff, and a social worker to target four main components during the prescribed intervention that included oral motor exercises, structured meals, behavioral interventions and caregiver training [61]. In this small study of three children with Goldenhar Syndrome, the authors reported a 36% increase in acceptance of food and drink, 16% decrease in the incidence of expulsion during mealtime, approximately 88% increase in mouth clean, and a 56% decrease in inappropriate mealtime behaviors following discharge from this multidisciplinary feeding program [61].

Trabi et al. retrospectively examined the impact of a program that utilized hunger promotion by reduction of tube feeding in a supervised setting, including an intensive, noninvasive monitoring of the child’s medical condition and full support of the child’s capacity for autonomous food exploration and self-regulated intake on the time to wean from tube feeding in 221 long-term feeding tube dependent participants [79]. The program included occupational, speech, and psychodynamic therapy in addition to nutritional counseling for the participants and their families. Following treatment, approximately 92% of the subjects weaned completely from tube feeding and none of the subjects returned to tube feeding after completing treatment [79].

Mckirdy et al. reviewed oral intake for two subjects enrolled in school-based interdisciplinary intervention programs with a team of professionals that consisted of school based speech therapists, consultative medical speech therapists, classroom teachers and aides, physical therapists, occupational therapists, school nurses, program administrators, and caregivers [71]. At the end of treatment, one participant consumed all medications and liquids by mouth but remained on a self-restricted diet. The other participant achieved 100% oral intake and removal of nasogastric tube following treatment [71].

Shaw et al. retrospectively examined the oral intake of solid foods, weight gain, duration of meals, and percentage of self-feedings in a six-year-old boy with Down syndrome following participation in a program that combined oral motor techniques (including the Beckman Oral Motor Program) speech therapy, and occupational therapy with behavior modification including reinforcement and punishment as well as stimulant medication [76]. At discharge, the subject was reportedly eating thick soup-based mixtures with some texture added from mashed vegetables, meats and noodles, and yogurt with fruit pieces. There was an increase in lingual range of motion, but oral feeding skill was still limited to munching. Variety and duration of intake improved, and the subject was able to adequately maintain a healthy weight without the use of a high calorie supplement. At a six-month follow up, weight had increased in accordance with developmental expectations [76].

Table 6

Five-phase model of clinical-outcome research in speech-language pathology

Level of evidence	Purpose	Examples
Phase 1	To determine if a therapeutic effect exists and determine its magnitude	Case Studies Single-subject studies Pre and post studies of small groups Retrospective analyses
Phase 2	To determine the total dimensions of a therapeutic effect and implement the necessary steps for clinical trial	Single-subject studies to test specific hypotheses Small group within-subject studies Case-control Small-group cohort-control Estimates effect size
Phase 3	To test efficacy of treatment strategies in clinical trials (establish internal validity)	Parallel-group designs
Phase 4	To determine the effect of a treatment strategy in clinical practice (establish external validity)	Pre and post studies of larger cohorts Parallel-groups Some hypothesis driven single-subject studies
Phase 5	To establish the benefits of the treatment strategy for specific populations and determine the cost of that treatment	Cost effectiveness studies Cost benefit analyses

Adapted from Robey [81].

Laud et al. retrospectively reviewed a variety of dependent variables including acceptance, refusal behaviors, negative vocalizations, grams consumed, caregiver measures, food variety inventories, and caregiver satisfaction scores in a large sample of subjects ( $n=$ 46) with diagnoses of ASD with or without co-occuring morbidity [68]. Behavioral and oral motor therapy implemented by an occupational therapist and a speech therapist included systematic meal sessions with individualized behavior protocols involving antecedent and consequence manipulation. Oral motor therapy was conducted to determine safety while eating and to perform eating and non-eating oral motor exercises. Significant differences were found in percentages of acceptance, refusal behaviors, negative vocalizations, and grams consumed from admission to discharge. Specifically, acceptance, refusal behaviors, and grams consumed increased significantly while negative vocalizations significantly decreased from admission to discharge. At three years following discharge, only 63% of the original sample could be contacted and 100% of those who could be contacted reported a significant increase in food acceptance and extinction of the structured therapeutic protocol [68].

3.12 Study quality

Clinicians providing treatment to children with feeding and/or swallowing disorders should be aware of the levels of evidence available to support the various treatment modalities used in their recommended programs. The search strategy utilized for the current EBSR did not reveal any published literature that examined the effects of sensory therapies (exercises and activities that are designed to promote sensory integration such as using a brushing or vibration protocol, a weighted vest, or deep pressure to the joints) or pharmaceutical therapies (including any pharmaceutical agent prescribed by a physician for the documented purpose of improving volume of oral intake) applied singularly (clinical questions two and four) in a therapeutic environment on functional feeding outcomes (i.e. volume of intake, variety of intake, days to wean tube feeding, weight gain, or growth). This is not to imply that there is not clinical or anecdotal evidence to support the use of these modalities to improve functional feeding outcomes, but this EBSR establishes that the current level of evidence to support these types of treatments applied singularly to treat pediatric feeding and/or swallowing problems is phase I research (least robust level of evidence, see Table 6 for hierarchical levels of evidence) [81]. All of the quality features are summarized in Table 2 for each article included in this EBSR.

Clinical question one identified one published article that addressed the effect of oral motor interventions on functional oral feeding outcomes in children. [13, 21]. Although it was designed as a randomized controlled trial it was underpowered and suffered from significant loss of subjects to follow-up, making post treatment assessments impossible for nine subjects (only 55% of the subjects had both pre and post treatment analysis). As a result, the authors did not find significant differences between the control and experimental groups on the outcomes of weight gain or change in oral motor function and this study does not offer any direct support for oral motor therapy to benefit functional oral feeding outcomes. Based on the results of this EBSR, oral motor therapy applications are phase I research. Significant limitations of this study included inadequately described study protocol, $p$ values neither reported nor calculable, and no effect sizes or confidence intervals reported or calculable (see Table 2) [21].

Of the 37 articles that addressed clinical question three, the majority (32/37) were determined to be phase one research reporting either 0/8 [27, 28, 37, 44], 1/8 [23, 24, 29, 30, 31, 36, 46, 48, 49, 55], 2/8 [26, 32, 33, 34, 35, 40, 41, 42, 43, 47, 51, 52, 54, 56, 57, 58], or 3/8 [50, 53] quality indicators. Limitations of the articles determined to be phase one research included small sample numbers (single subject or case controlled trials), with inadequately described protocols (insufficient detail for replication), utilizing convenience samples with inconsistently described evidence of treatment fidelity, and inconsistent reports of significance precision (effect size). The five remaining articles relevant to clinical question three were randomized controlled trials and determined to be phase four research [22, 25, 38, 39, 45]. None of the randomized controlled trials provided 8/8 quality markers, please see Table 2 for complete summary of included features.

The 23 studies identified by this EBSR relevant to clinical question five can be described as phase one research with all of them reporting two or less quality markers (seven studies reported 0/8 quality markers [18, 60, 71, 76, 77, 78, 79], five reported only 1/8 quality markers [59, 61, 62, 64, 75], and 11 reported 2/8 quality markers [63, 65, 66, 67, 68, 69, 70, 72, 73, 74, 80]. Only three of the 23 articles relevant to mixed modality interventions reported on the statistical significance of their findings [63, 68, 73].

4. Discussion

This EBSR summarized the published evidence to support common treatment modalities for children with feeding and/or swallowing disorders. The results bring to light the lack of published evidence to support the singular use of oral motor interventions (one of the 61 included articles) [21], sensory-based, and pharmaceutical interventions (zero of the 61 included articles) to remediate pediatric feeding and/or swallowing problems. Additionally, it highlights for the reader the moderate amounts of published evidence to support the use of behavioral interventions (37 of the 61 included articles) [22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], the use of mixed method interventions (23 of the 61 included articles) [18, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80] and the various levels of evidence available to support the use of behavioral (32/27, 86% were found to be phase one research) and combined treatment options (all phase one research with extensive methodological limitations [18, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80]) for children with feeding and/or swallowing disorders. The clinical utility of the findings from question one (oral motor interventions) [21] and question five (mixed method intervention strategies) [18, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 79, 80] is limited due to the limitations of the methodologies employed in these studies and the primarily descriptive data reported. The lack of scientific evidence available on the singular use of oral motor interventions, sensory-based interventions, and pharmaceutical interventions does not allow clinicians to draw any valid conclusions about the efficacy of these methods when applied for improving functional feeding outcomes in pediatric populations. However, the moderate amount of published, phase one evidence on mixed method interventions does call for further investigation of these strategies with greater scientific rigor to establish the full dimensions of the therapeutic effect and ultimately determine the internal and external validity of these treatment options as an initial therapeutic effect has been established.

From this EBSR, clinicians recognize the importance of behavioral therapy techniques for remediating feeding disorders and must ensure that members of the multidisciplinary team are trained in the appropriate application of the various behavioral therapy techniques. Sixty of the 61 articles identified utilized behavioral therapy techniques within the prescribed therapeutic plan and reported various levels of improvement in functional feeding outcomes including increased dietary intake, variety, and reduction in inappropriate mealtime behaviors [18, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 80]. Five of the sixty identified articles that reported on behavioral feeding therapy approaches were randomized controlled trials and met the criteria of phase four evidence [22, 25, 38, 39, 45].

One study randomized young patients with cystic fibrosis to either a wait list (control group) or a behavioral and nutrition intervention (experimental group) and found that the combination of behavioral and nutrition intervention led to an increase in energy intake as compared to wait list [45]. Two [22, 39] of those five articles reported on a specific program (Hassle Free Mealtimes Triple P) of behavioral parent training to remediate difficult mealtime behaviors in children and improve parent variables related to those behaviors in otherwise healthy, typically developing children [22, 39]. One utilized the full program [22] and the other utilized an abbreviated version of the program [39]. Both reported significant improvements in mealtime behaviors following treatment with the parent behavioral training program [22, 39].

The remaining two randomized controlled trials identified were designed to compare the effectiveness of two different approaches one to wean children from tube feedings [25] and the other to increase dietary intake, variety, and decrease problematic mealtime behaviors [38]. Benoit and colleagues compared the effectiveness of nutritional training with nutritional training with the added component of extinction and found that the nutritional training with the extinction component was significantly more effective in weaning the use of tube feeding dependency in children with a mean age of 17.2 months [25]. Marshall et al. compared the effectiveness of systematic desensitization and operant conditioning approaches to improving dietary intake, variety and decreasing problematic mealtime behaviors and found no significant differences in any of the primary outcomes between participants in the two groups [38]. Marshall and colleagues well designed randomized controlled trial highlights that tenants of both behavioral approaches (operant conditioning and systematic desensitization) were effective in significantly improving dietary intake and variety, and reducing problematic mealtime behaviors when administered to a protocol by reliably trained professionals [38].

Across all of the studies, methodologies for evaluating the treatment modality in question were inconsistent and there was wide variability among patient populations studied (including children with ASD, medically and non-medically complex histories, genetic syndromes, etc.…) which in combination with the relatively small numbers of subjects enrolled in the majority of the studies included in this review, limits the generalizability of the results to broad clinical populations. Clinicians are aware that there are three tenants to evidence based practice: external scientific evidence, clinical expertise/opinion, and client/patient/caregiver perspectives and desires. Whenever possible, clinical care should be based on the best available evidence, but when there is insufficient quantity and/or quality of evidence available in the literature, clinicians must determine what is best practice based on the limited available evidence, clinical opinion, and the needs/desires of their patients and their families. Further, clinicians must focus on functional feeding outcomes for their clients, document client progress, and adjust treatment modalities used to fit the individual needs of each client.

Limitations of the current EBSR included the inclusion criteria of exclusively peer-reviewed literature published only in English. This design feature may have contributed to a publication bias in this EBSR. Due to the limited resources available for this EBSR, limiting inclusion to only articles published in English was a necessity; however, the authors acknowledge that this may have introduced bias and ultimately limited the number of articles identified. Additionally, reliability of reviewers could not be determined, but the three reviewers independently reviewed all of the articles before inclusion and then used discussion and consensus to resolve any discrepenancies.

5. Conclusion

This EBSR contributes to clinicians’ knowledge of the published literature on the use of various treatment modalities available to treat pediatric feeding and swallowing disorders including oral motor, sensory, behavioral, pharmaceutical, and mixed method intervention strategies. It is yet to be established whether or not oral motor, sensory, or pharmaceutical interventions applied singularly or in combination can be used to successfully aid in achieving functional feeding outcomes for pediatric patients. There is phase four research from randomized controlled trials that establishes the following:

A program that incorporates behavioral and nutrition intervention for young children with cystic fibrosis leads to greater increases in energy intake as compared to usual care [45].

Parent-training programs are successful in reducing problematic feeding behaviors in otherwise healthy and typically developing children [22, 39].

Nutritional intervention in combination with behavioral therapy (extinction technique) is more effective in weaning young children from tube feeding dependency than nutritional intervention alone [25].

Both systematic desensitization and operant conditioning behavioral therapy approaches are effective in improving dietary intake, variety, and decreasing problematic mealtime behaviors in children with ASD and those with non-medically complex histories when administered to a protocol by trained professionals [38].

Funding

There was no funding source for this study.