Feeding Difficulties Intervention in Medically Complex Children

Journal of Pediatric Gastroenterology and Nutrition Publish Ahead of Print
DOI: 10.1097/MPG.0000000000001683
Intervention for Feeding Difficulties in Children with a Complex Medical History:
A Randomized Clinical Trial
Jeanne Marshall, B.Sp.Path, Ph.D1
Rebecca J. Hill, B.App.Sci (Hons), Ph.D, R.Nutr2
Meagan Wallace, B.A.Sc, MSpPathSt1
Pamela Dodrill, B.Sp.Path (Hons), Ph.D1
1
Children’s Health Queensland, Lady Cilento Children’s Hospital, Brisbane, Australia
2
The Children's Nutrition Research Centre, UQ Child Health Research Centre, The
University of Queensland, Brisbane, Australia
Correspondence:
Jeanne Marshall
Speech Pathologist (Clinical Lead)
Level 7a
Lady Cilento Children’s Hospital
Raymond Terrace
South Brisbane, Australia, 4101
Phone: +617 3068 5286
Email: [email protected]
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
Running head: Intervention for feeding difficulties
Trial registration:
This trial was registered on the Australia New Zealand Clinical Trials Registry
http://www.anzctr.org.au Ref #ACTRN12611000956909
Declaration of funding sources:
Queensland Children’s Medical Research Institute
The University of Queensland
Children’s Health Queensland
The authors have no conflict of interest to declare.
Number of figures: 2
Number of tables: 3
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 (www.jpgn.org).
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
Abstract
Objective: This study aimed to compare outcomes of different multidisciplinary feeding therapy
approaches in children with feeding difficulties.
Methods: Children 2-6 years with feeding difficulties and a medically complex history (MC)
were recruited. Children with feeding difficulties and a non-medically complex history (NMC)
were included as a comparison group. Participants attended a clinical assessment, and eligible
participants were randomized to receive targeted feeding intervention incorporating either operant
conditioning or systematic desensitization. Parents could elect to receive intervention in an
intensive (10 sessions in a week) or weekly (10 sessions over 10 weeks) format. Both groups
received immersive parent training. A review was completed three months post-intervention.
Results: In total, 98 participants were eligible to participate (MC n=43; NMC n=55). Data from
20 children from the MC group (47%) and 41 children from the NMC group (75%) were included
in the final analysis. Clinically significant improvements were observed following both arms of
therapy, consistent with previous research. Parents of children in the MC arm were significantly
more likely to elect for intensive intervention than weekly (MC=12/20, 60%; 12/41, 29%;
p=0.02).
Conclusions: Both therapy protocols were considered clinically effective. The difference in
attrition rates between the etiological groups suggests primary differences in how service delivery
should be managed. Progress for the medically complex child may be slower while medical
issues are stabilized, or while the focus for parents shifts to other developmental areas. In
planning services for a medically complex group, therefore, it is essential that consideration be
given to medical and family needs.
Key words: Feeding difficulties; intervention; medically complex; operant conditioning;
systematic desensitization
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
What is known:
 Children with a medically complex history frequently present with feeding
difficulties
 There is a paucity of literature evaluating or comparing intervention options for
this group of children
What is new:
 Targeted feeding therapy incorporating either systematic desensitization or
operant conditioning techniques resulted in improvements in dietary variety,
difficult mealtime behaviors, and parental stress for children with a medically
complex history
 Significantly higher attrition rates in the medically complex group versus the nonmedically complex group suggest primary differences in how service delivery for
feeding difficulties should be managed in this group
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
Introduction
Feeding difficulties are common in children with a complex medical history, with
the incidence reported as between 40 and 70% (1). There are several mechanisms that
may underlie the development of feeding difficulties for children in this group. Children
with complex medical conditions may have differences in anatomy, muscle strength, and
swallow coordination that limit the development of typical feeding (1). The onset of
feeding difficulties in children with a complex medical history may be influenced by the
application of unpleasant medical interventions necessary to sustain life (2, 3) or the
repeated association of eating with pain or discomfort (4, 5). In addition, a lack of
experience with eating during a child’s early life as a result of illness or alternative
feeding may contribute to the development of delayed oral motor skills (4, 6), or oralsensory based disorders (3), which further exacerbate limitations in dietary variety.
Regardless of any underlying mechanism, in many cases, feeding difficulties persist
beyond the resolution of the original medical condition that precipitated them, and
families of children with complex medical needs report feeding difficulties to be one of
their most common chronic health problems (7).
With regards to treatment, much of the research is retrospective and focused on
small subgroups of children (8), which makes generalizability difficult. Intervention
based on operant conditioning (OC) is commonly used (9, 10), and can involve a range of
OC techniques broadly including reinforcement, punishment, and escape extinction. OC
intervention generally involves a prompt-and-reward structure, and is externally driven
(e.g. the child needs to take a bite to receive the reward) (11). Intervention based on
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
systematic desensitization (SysD) is emerging in the literature as a newer treatment style
for feeding difficulties (8, 11), and involves gradual exposure to a feared stimulus within
a relaxing environment. SysD is typically play-based, and internally driven (e.g. the child
is not instructed to take a bite, and only does so out of their own volition). The primary
contrast between these two interventions is that SysD is generally bottom-up therapy
(look, smell, touch, then taste and eat), whereas OC is generally top-down therapy
(starting at the level of taking a bite). Regardless of treatment style, much of the literature
advocates for the use of a multidisciplinary team for evaluation and management (3, 4, 8,
12) to ensure motor and sensory skill deficits impacting feeding are addressed, as well as
behavior and nutrition goals.
This study aimed to replicate a previously successful randomized clinical trial
(RCT) (11), but with children with a complex medical history (MC). Children with a nonmedically complex (NMC) history are included as a comparison group. This study
focused on children with feeding difficulties, and had three aims: (1) To examine the
baseline characteristics in MC vs. NMC participants; (2) To examine the outcomes of
therapy intervention for (a) MC vs. NMC participants overall; (b) OC versus SysD
intervention, and (c) intensive versus weekly therapy intensity dose; and (3) To examine
the parent satisfaction following access to a feeding therapy program.
Methods
See the Supplemental Digital Content, Methods, http://links.lww.com/MPG/B51, for the
complete Methods section.
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
Results
Participant demographics
Overall, 124 potential participants enquired during the study period and were
formally screened for eligibility (see Figure 2). Of these, 98 were eligible for initial
assessment (MC=43; NMC=55), 64 completed the full intervention course (MC=21;
NMC=43), and 61 complete datasets were available for analysis (MC=20; NMC=41).
Across the MC arm, participants with a variety of complex medical backgrounds were
included, and these were categorized according to primary medical specialty accessed
(Ex-premature n=8; Oncology n=1; Gastroenterology n=2; Respiratory n=4; Genetic n=1;
Cardiac n=2; Cleft palate n=2). Total attrition for the MC arm was significantly greater
than that observed in the NMC arm (MC=23/43, 53%; NMC=14/55, 25%; p<0.01).
Reasons for attrition across both arms included other family commitments, family
tragedy, the family not wishing for their child to be randomized, location of the study,
being lost to follow up, being underweight, and receiving an ASD diagnosis during the
course of the study. A reason for attrition that was characteristic only of the MC arm was
the presence of new or ongoing medical issues. There were three participants in the NMC
arm that indicated they no longer required support after the assessment phase of the
study.
Insert Figure 2 near here
There were two participants that started a course of intervention (one randomized
to SysD, one randomized to OC), but there were issues with therapy suitability, and these
participants were switched to the therapy protocol of the opposing arm. In both cases,
participants had not demonstrated any therapy response by the fourth therapy session,
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
parents were demonstrating increasing feeding-related stress, and it was determined by
the treating clinicians that the randomized course of treatment had the potential to worsen
mealtime behaviours. These cases were included by using an intention-to-treat analysis as
per study design.
Comparison of MC and NMC groups at baseline
Demographic and baseline assessment data for both groups are presented in Table
1. The groups presented with many similar characteristics. The MC group more
commonly presented with a history of tube feeding (MC=10/20, 50%; NMC=2/41, 5%;
p<0.01), feeding therapy (MC=12/20, 60%; NMC=12/41, 29%; p=0.04), dietetics input
(MC=12/20; NMC=12/41; p=0.04), and hospitalizations (MC=1.6±1.5; NMC=0.5±0.9;
p<0.01). The MC group were eating significantly more protein rich foods (MC=8.2±3.5;
NMC=5.7±4.2; p=0.03), and more unprocessed fruits and vegetables (MC=5.9±7.1;
NMC=2.5±2.8; p=0.05) than the NMC group before intervention. A high proportion of
children across both arms presented with oral motor delays (n=42/58; 72%) and/ or oral
sensory sensitivity (n=35/59; 59%), which is consistent with a previous paper in the area
(13).
Insert Table 1 near here
Organization of MC and NMC groups across different arms and intervention
intensities can be observed in Table 2. Randomization across the therapy arms was
evenly distributed. Parents of children in the MC arm were significantly more likely to
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
elect for intensive intervention than weekly intervention (MC=12/20, 60%; 12/41, 29%;
p=0.02).
Outcomes at 3 month follow-up
Overall, for this group of children, there were statistically and clinically
significant changes observed in the reduction of difficult mealtime behaviors, reduction
of parental stress ratings, and increased dietary variety after intervention (see Table 3).
Review of effect size calculations revealed the changes in the frequency of reported
difficult mealtime behaviors (p<0.01; d=-0.87) and the number of protein-rich foods
consumed (p<0.01; d=0.91) were large. There were small to moderate effects
demonstrated across the majority of other variables. All participants were within an
acceptable range for height, weight, and BMI before, during, and after therapy. Post-hoc
analysis was completed to investigate the differences in outcomes between the MC and
NMC groups. This revealed that there were no significant differences in the degree of
change between the MC and NMC groups, suggesting that both groups benefited from
the interventions provided.
Using texture analysis as a novel method for measuring dietary change (see
Figure 1)(14), there was a significant increase observed with regards to the total
percentage of soft mechanical foods (i.e. foods that have to be chewed) consumed across
the groups as a whole (pre=34.7%; post=38.6%; p=0.02). Subsequently, overall,
participants were observed to be consuming less minced/moist foods and dissolvable
foods to compensate (i.e. foods that require no active chewing). Further analyses revealed
that, compared to the NMC group, the MC group made significantly more change in their
reduction of caloric intake via fluids (pre=20.6%; post=16.9%; p=0.04), and were eating
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
more soft mechanical foods instead, (pre=36.4%; post=42.2%; p=0.24), which was
considered to be clinically significant. Although not statistically significant, it was
considered clinically significant that children identified with moderate oral motor
difficulties at baseline were consuming more liquids, purees, and minced/moist textures,
and less soft mechanical foods than children with typical or mild oral motor skills
difficulties.
Post-hoc analysis of outcomes for the OC vs SysD groups revealed no significant
differences in the data between groups. Finally, when comparing the intensive arm to the
weekly arm, there were no significant differences in outcomes between the groups.
Parent satisfaction
Overall, parents reported significant improvements in satisfaction ratings for their
children’s dietary variety (mean pre=1.62; immediately post =2.8; mean 3 month followup=2.9; p<0.01), overall nutrition (mean pre=2.5; immediately post=3.1; mean 3 month
follow-up=3.2; p<0.01), and mealtime behavior (mean pre=2.3; immediately post=3.4;
mean 3 month follow-up=3.4; p<0.01). Post-hoc analyses demonstrated that
improvement in satisfaction ratings was significant between pre- and immediate posttreatment measures; satisfaction ratings taken immediately post-treatment and at the 3month review remained the same. Further analyses revealed no differences in the increase
of satisfaction ratings between the MC and NMC groups, OC and SysD interventions,
and weekly and intensive intervention frequencies.
The majority of parents reported that the feeding therapy program assisted them to
better understand their child’s feeding difficulties (n=57; 93%), and for most parents, this
was more than they expected (n=39, 64%). Overall, 100% of parents surveyed (n=60, 1
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
incomplete data) reported that they would recommend the interventions offered to other
parents, which is suggestive of a high degree of confidence in the program. Many parents
(n=45; 74%) reported that they felt their child would benefit from another block of
therapy. It was considered clinically significant that more families that had attended an
intensive treatment block vs. a weekly block requested another block of therapy (weekly
n=25; 68%; intensive n=20; 87%; p=0.08).
Fidelity to treatment protocol
Total compliance with the intervention protocol was 88% across all participants
(n=61). The MC group achieved marginally lower fidelity scores than the NMC group
(MC group=17.7/21, 84%; NMC group=18.8/21, 90%; p=0.11), but this difference was
not statistically significant. The most common feature where the intervention protocol
was not complied with was session attendance. Participants were always offered 10
sessions, but for various reasons, may not have attended all 10, and prior to study
commencement it was decided that a minimum of 7 sessions would be adequate for data
inclusion. The median number of sessions attended was 9. The total mean number of
foods offered during the course of treatment across both treatment arms was 30.0 as per
the protocol (OC=30.3±2.04 foods; SysD=29.6 ±1.2 foods; p=0.13).
Discussion
This is the first known randomized clinical trial to compare outcomes of two
styles of feeding therapy for medically complex children. In a similar manner to the
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
previous study in this series (11), statistically and clinically significant favorable changes
to outcome measures for children receiving either intervention were observed.
Medically complex vs. non-medically complex
Although there were no significant differences in outcomes between the MC and
NMC groups, there were a number of features of difference that prompt discussion. One
such feature was the difference in attrition between the groups, where the MC group had
a significantly higher degree of attrition over the course of the study. While this occurred
for a range of practical reasons across both groups, participants in the MC arm were more
likely to withdraw as a result of illness. Another feature of disparity between the MC and
NMC group was the difference in fidelity scores, where the MC group had lower scores
overall. This reflects the fact that the intervention protocol was somewhat more difficult
to follow with this group of children. Finally, it was observed that parents of children in
the MC arm were significantly more likely to opt for an intensive treatment block
compared to parents of children in the NMC arm. The reasons for this are unknown, but
they may have been related to the anticipated potential for illness, or the desire for quick
resolution of feeding difficulties. It could be postulated that families familiar with a
medical-style model of treatment (e.g. where the child is an inpatient) may have
perceived this style to be more successful as a “quick-fix”. It is proposed that although
intervention itself may be similar across different groups of children with feeding
difficulties, practical differences in management for MC children should be taken into
consideration in planning service delivery. Direct engagement with families to determine
the reasons behind service delivery choices would be a valuable direction for research in
developing models of care for MC children.
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
Operant conditioning vs. systematic desensitization
This study replicates previous research findings (11) by demonstrating
statistically and clinically significant changes to outcome measures for children receiving
targeted feeding therapy incorporating either OC or SysD techniques. This suggests that,
when delivered to a protocol, with consideration of the sensory motor skills of the child,
and with the inclusion of parent training, operant conditioning or systematic
desensitization approaches can be successful forms of treatment for feeding difficulties.
In two cases, however, there were significant issues with suitability of the therapy option
the participants were randomized to, and participants were switched to the opposing arm.
Whilst it was not an unexpected outcome for the study, the need to switch intervention
arms for two of the participants is recognized as a limitation. The aims of providing
intervention were to improve mealtime behaviours, increase dietary variety and decrease
parental stress and, in these two cases, it was determined that these aims were not being
achieved. These cases were included in an intention-to-treat analysis. In this study, we
randomized participants to determine if treatments were equivalent overall. However, in
clinical practice, it is likely that each intervention style has features that may suit some
children and families better than the other. It is suggested that a ‘recipe approach’ does
not necessarily suit all families, and that some flexibility and individuality should be
applied where required.” It also stresses the importance that health facilities should have
a variety of intervention options available for families.
Oral motor skills
Texture analysis of the 3-day food diaries demonstrated that participants with
moderate oral motor impairments in this study were consuming more volumes of foods
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
that were easier to eat, and that variety of textures improved with intervention. Whilst
there are some standardized assessments available to assist with identification of severe
oral motor impairment (e.g. (15)), there is a paucity of quality oral motor assessments
available for children with mild or moderate impairments. Although the assessment used
in this study was novel, the association of easier-to-eat textures with poorer oral motor
skill, and the demonstration of improvement, suggests that functional assessment through
examination of the diet may yield some valuable information in this area. The
development of a standardized assessment with greater sensitivity for mild/moderate oral
motor impairments, and one that could accurately capture change is an area that warrants
further research.
Parent training
Parental involvement has been shown to improve maintenance and generalization
of therapy skills to the home environment (16). Improvements in parent satisfaction
ratings for this study were demonstrated, and it is suggested that these may have been as
a result of the provision of a parent-training program that involved a range of
opportunities for skills immersion and feedback. It is noteworthy though that many
parents indicated that they felt that their child would still benefit from further therapy
intervention. Additionally, parents reported little change in satisfaction between
completion of the therapy block and the 3-month follow-up session, which suggested that
they didn’t feel there was further improvement once they were no longer receiving active
input for their child. Further research is required to determine what teaching styles and
treatment doses best support parents in feeling confident to manage feeding difficulties in
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
their children. Ideally, the goal of treatment is to empower parents to manage their child’s
feeding issues, as opposed to attending prolonged therapy blocks.
Limitations
Some limitations are acknowledged in this study. One of these was the high level
of attrition observed in the MC group, which may have prevented the ability to fully
characterize feeding difficulties or response to treatment in this group of children. While
every care was taken to prevent attrition, the high level in itself does serve to act as a
characteristic of the medically complex group of children. Some allowance for competing
needs in this group should, therefore, be considered in planning service delivery.
It is also recognized that the design for this study excluded children with severe
feeding difficulties, and, as a result, there are limited conclusions that can be drawn about
treatment in this population of children. Increasing severity of feeding difficulties,
including those children who are tube-fed or those with aspiration risk, necessitates the
support of close medical support, which was not immediately available in our outpatient
model. Increasing severity of feeding difficulties also requires therapy to be more
specialized, which makes development and adherence to a therapy protocol very difficult.
Finally, it is acknowledged that there was no control group used in this study,
which would have provided information about whether participants would have made the
same improvements without active treatment. After much consideration and consumer
feedback during the design of the study, it was decided that use of a control group would
not be feasible, as many parents were unwilling to consent to participation if there was a
chance they would not receive treatment, or indicated they would cease participation if
they were randomized to a control condition where no active therapy was provided. In
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
addition, time limitations and funding of the study unfortunately did not permit a time-lag
design. We considered it relevant that participants accessing the study had presented with
chronic feeding difficulties, and parents were concerned enough to volunteer and commit
to the time-intensive program. As a result of these factors, we hypothesize that feeding
difficulties in our cohort would have been unlikely to resolve without intervention,
although we acknowledge that this conclusion cannot truly be drawn without a control
group.
Conclusion
This study provides valuable information with regards to treatment for feeding
difficulties in children with a complex medical history. Medically complex and nonmedically complex children demonstrated similar improvements in treatment outcomes
after structured multi-disciplinary therapy input. However, differences in attrition and
service delivery choices between groups are a reflection of other issues in the family life
of the medically complex child, and are an important consideration in planning treatment
for this group. This randomized clinical trial builds on previous research, and provides
further evidence that, when delivered to a protocol by an experienced multidisciplinary
team and accompanied by parent training, the use of targeted therapy approaches
incorporating either operant conditioning or systematic desensitization may be successful
in the treatment of children with mild/ moderate feeding difficulties.
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
References
1. Rudolph CD, Link DT. Feeding disorders in infants and children. Pediatr Clin North
Am. 2002;49(1):97-112.
2. Douglas JE, Bryon M. Interview data on severe behavioural eating difficulties in
young children. Arch Dis Child. 1996;75(4):304-8.
3. Rommel N, De Meyer AM, Feenstra L, Veereman-Wauters G. The complexity of
feeding problems in 700 infants and young children presenting to a tertiary care
institution. J Pediatr Gastroenterol Nutr. 2003;37(1):75-84.
4. Cooper-Brown L, Copeland S, Dailey S, Downey D, Petersen MC, Stimson C, et al.
Feeding and swallowing dysfunction in genetic syndromes. Dev Disabil Res Rev.
2008;14(2):147-57.
5. Zangen T, Ciarla C, Zangen S, Di Lorenzo C, Flores AF, Cocjin J, et al.
Gastrointestinal motility and sensory abnormalities may contribute to food refusal in
medically fragile toddlers. J Pediatr Gastroenterol Nutr. 2003;37(3):287-93.
6. Edwards S, Davis AM, Bruce A, Mousa H, Lyman B, Cocjin J, et al. Caring for tubefed children: a review of management, tube weaning, and emotional considerations.
JPEN J Parenter Enteral Nutr. 2016 Jul;40(5):616-22.
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
7. Kuo DZ, Berry JG, Glader L, Morin MJ, Johaningsmeir S, Gordon J. Health services
and health care needs fulfilled by structured clinical programs for children with medical
complexity. J Pediatr. 2016;169:291-6.e1.
8. Davis AM, Bruce A, Cocjin J, Mousa H, Hyman P. Empirically supported treatments
for feeding difficulties in young children. Curr Gastroenterol Rep. 2010;12(3):189-94.
9. Sharp WG, Stubbs KH, Adams H, Wells BM, Lesack RS, Criado KK, et al. Intensive,
manual-based intervention for pediatric feeding disorders: results from a randomized
pilot trial. J Pediatr Gastroenterol Nutr. 2016;62(4):658-63.
10. Marshall J, Ware R, Ziviani J, Hill RJ, Dodrill P. Efficacy of interventions to improve
feeding difficulties in children with autism spectrum disorders: a systematic review and
meta-analysis. Child Care Health Dev. 2015;41(2):278-302.
11. Marshall J, Hill RJ, Ware RS, Ziviani J, Dodrill P. Multidisciplinary intervention for
childhood feeding difficulties. J Pediatr Gastroenterol Nutr. 2015;60(5):680-7.
12. Conforti A, Valfre L, Falbo M, Bagolan P, Cerchiari A. Feeding and swallowing
disorders in esophageal atresia patients: a review of a critical issue. Eur J Pediatr Surg.
2015;25(4):318-25.
13. Marshall J, Hill RJ, Ware RS, Ziviani J, Dodrill P. Clinical Characteristics of 2
Groups of Children With Feeding Difficulties. J Pediatr Gastroenterol Nutr.
2016;62(1):161-8.
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
14. Dodrill P. Feeding problems and oropharyngeal dysphagia in children. J
Gastroenterol Hepatol Res. 2014;3(5):1055-60.
15. Reilly S, Skuse D, Wolke D. Schedule for oral motor assessment: Administration
manual. London: Whurr Publishers Ltd.; 2000.
16. Lerma DC, Swiezy N, Perkins-Parks S, Roane HS. Skill acquisition in parents of
children with developmental disabilities: interaction between skill type and instructional
format. Res Dev Disabil. 2000;21(3):183-96.
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
Figure 1. Food textures
Pureed foods
(e.g. rice cereal, yoghurt)
Mashed foods
(e.g. mashed vegetable,
avocado)
Minced/ moist textures
(e.g. baked beans in sauce)
Oral motor skill required to consume texture
Bolus taken from a spoon;
Tongue moves in a forward-backward plane of movement
Bolus taken from a spoon, fork, or fingers;
Tongue moves in a forward-backward plane of movement; upwards tongue pressure is used to compress
foods between the tongue and palate
Bolus is usually taken from a spoon or fork; components of food are fluid or pureed, and tongue moves in a
forward-backward plane of movement; other components of food need to be masticated by the teeth, and
tongue lateralization is required to move food onto the chewing surfaces.
Bolus is cut up or broken then taken from a fork or fingers;
Sideways tongue movement (tongue lateralization) is required to move food onto the chewing surfaces
Soft mechanicals*
(e.g. cooked chicken, pasta,
cheese)
Bolus is cut up or a bite is taken; the bolus is usually taken from fingers or possibly a fork;
Hard mechanicals*
(e.g. beef steak, raw carrot) Sideways tongue movement (tongue lateralization) is required to move food onto the chewing surfaces
*‘Mechanical’ food is another term for food that requires chewing
Figure adapted from
Dodrill P. Treatment of Feeding and Swallowing Difficulties in Infants and Children. In: Groher M, Crary MA, editors. Dysphagia: Clinical
Management in Adults and Children. 2 ed. Missouri, U.S.A.: Mosby, Inc.; 2016.
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
Enrollment
Assessed for eligibility (n=124)
Excluded from this particular study
(n=26)
 ASD diagnosis (n=10)

Outside age bracket (n=6)
 Tube-fed (n=5)

Underweight (n=4)
 > 2 allergies (n=1)
Eligible (n=98)
Assessment
Medically complex arm (n=43)
 Received assessment (n=30)
Non-medically complex arm (n=55)
 Received assessment (n=45)
 Did not receive assessment (n=13)
 Did not receive assessment (n=10)
 Lost to follow-up (n=6)
 Family too busy (n=3)
 New/ ongoing medical issues (n=3)
 No longer required support (n=3)
 Study too far away (n=2)
 Study too far away (n=2)
 Family too busy (n=1)
 Family tragedy (n=1)
 Family did not wish to be randomised (n=1)
 Family did not wish to be randomised (n=1)
Intervention
 Received intervention (n=21)
 Received intervention (n=43)
 Did not receive intervention (n=9)
 Did not receive intervention (n=2)
 Underweight (n=1)
 Lost to follow up (n=1)
 Family too busy (n=4)
 New/ ongoing medical issues (n= 2)
 Lost to follow-up (n=1)
 Family tragedy (n=1)
 Underweight (n=1)
Analysis
 Analyzed (n= 20)
 Analyzed (n= 41)
 Excluded from analysis (n=1)
 ASD diagnosis (n= 1)
 Excluded from analysis (n=2)
 ASD diagnosis (n= 1)
 Lost to follow up (n=1)
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
Table 1. Participant demographics at baseline
Male (n, %)
Only child (n, %)
Problems during birth or pregnancy (n, %)
Caesarean section birth (n, %)
History of feeding therapy (n, %)
Current or previous dietetics input (n, %)
History of tube feeding (n, %)
≥2 areas delayed (n, %)
Oral motor delay (n, %)
Definite difference in oral sensory sensitivity (n, %)
Age in months (m, SD)
Number of hospitalizations (m, SD)
Height z-score (m, SD)
Weight z-score (m, SD)
BMI z-score (m, SD)
ECBI intensity score (m, SD)
ECBI problem score (m, SD)
BPFAS frequency score (m, SD)
BPFAS problem score (m, SD)
PPFSQ parent stress score (m, SD)
PPFSQ child stress score (m, SD)
PSI total percentile (m, SD)
Total grains (m, SD)
Total proteins (m, SD)
Total unprocessed fruit and veg (m, SD)
Total empty calories (m, SD)
Total food count without empty calories (m, SD)
Total percentage of energy met (m, SD)
MC (n=20)
14 (70%)
14 (70%)
12 (60%)
7 (35%)
12 (60%)
12 (60%)
10 (50%)
6 (30%)
Typical=6 (33%); Mild=6 (33%);
Mod.=6 (33%)*
12 (63%)**
41.7 (±13.2)
1.6 (±1.5)
-0.02 (±1.2)**
0.06 (±1.2)
0.07 (±1.2)**
53.4 (±8.0)
52.1 (±9.7)
101.6 (±16.4)
20.3 (±6.7)
2.6 (±1.0)
1.7 (±0.8)
60.4 (±19.3)
5.6 (±3.3)
8.2 (±3.5)
5.9 (±7.1)
3.0 (±3.2)
23.8 (±11.3)
91.9 (±19.9)
NMC (n=41)
30 (73%)
31 (76%)
23 (56%)
15 (37%)
12 (29%)
12 (29%)
2 (5%)
12 (29%)
Typical=10 (25%); Mild=12 (30%);
Mod.=18 (45%)^
23 (58%)^
47.9 (±11.2)
0.5 (±0.9)
0.3 (±1.0)
0.23 (±0.8)
0.04 (±0.7)
56.6 (±6.3)
52.2 (±8.5)
103.3 (±14.9)
18.6 (±6.9)
2.3 (±1.0)
1.7 (±1.0)
59.0 (±20.5)
6.0 (±3.2)
5.7 (±4.2)
2.5 (±2.8)
4.5 (±3.8)
22.2 (±10.3)
91.0 (±14.6)
p-value
1.00
0.88
0.99
1.00
0.04
0.04
<0.01
1.00
0.68
0.90
0.06
<0.01
0.24
0.51
0.91
0.09
0.94
0.68
0.36
0.24
0.90
0.81
0.65
0.03
0.05
0.11
0.59
0.85
*n=18; **n=19; ^n=40
Mod.=Moderate delay; BMI=Body Mass Index; ECBI=Eyberg Child Behavior Inventory (27); BPFAS=Behavioral Pediatrics Feeding Assessment Scale (22);
PPSFQ=Parent Perceived Feeding Stress Questionnaire (30); PSI=Parenting Stress Index (28, 29)
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
Table 2. Randomization and intensity across MC and NMC arms
Medically complex arm (n=20)
Operant conditioning
Systematic desensitization
4
4
Weekly intervention
7
5
Intensive intervention
11
9
TOTAL
Non-medically complex arm (n=41)
13
16
Weekly intervention
7
5
Intensive intervention
20
21
TOTAL
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
TOTAL
8
12
20
29
12
41
Table 3. Changes to outcome measures at 3-month follow-up
Variable
Pre-mean
Post-mean (±SD)
(±SD)
Height z-score
0.22 (±1.06)
0.17 (±1.08)
Weight z-score
0.17 (±0.93)
0.17 (±0.97)
BMI z-score
0.05 (±0.84)
0.12 (±0.95)
ECBI intensity T-score
55.6 (±7.0)
53.7 (±8.2)
ECBI problem T-score
52.2 (±8.8)
49.4 (±9.4)
BPFAS total frequency score
102.7 (±15.3)
88.7 (±16.9)
BPFAS total problem score
19.2 (±6.8)
13.2 (±7.8)
PPSFQ parent stress rating
2.4 (±1.0)
1.7 (±1.0)
PPSFQ child stress rating
1.7 (±1.0)
1.3 (±0.8)
PSI percentile
59.4 (±19.9)
55.5 (±27.3)
Total grains
5.9 (±3.2)
7.1 (±3.7)
Total proteins
6.5 (±4.1)
10.6 (±4.9)
Total unprocessed fruit and vegetables
3.6 (±4.9)
5.4 (±6.1)
Total empty calories
4.0 (±3.6)
5.2 (±3.8)
Total food count without empty calories
22.7 (±10.6)
29.1 (±12.5)
Total percentage energy intake
91.3 (±16.4)
99.4 (±22.2)
Total percentage fluids
18.1 (±11.7)
18.0 (±12.2)
Total percentage purees
13.7 (±13.3)
13.1 (±13.1)
Total percentage minced/moist
8.8 (±9.9)
7.8 (±7.7)
Total percentage soft mechanicals
34.9 (±16.3)
38.6 (±16.4)
Total percentage hard mechanicals
15.8 (±12.5)
15.5 (±10.6)
Total percentage dissolvables
8.4 (±9.3)
7.0 (±7.4)
Mean difference
(95% CI for difference)
-0.06 (-0.44, 0.33)
0.00 (-0.34, 0.34)
0.07 (-0.25, 0.39)
-1.8 (-4.6, 0.9)
-2.7 (-6.0, 0.5)
-14.0 (-19.8, -8.3)
-5.9 (-8.6, -3.3)
-0.7 (-1.1, -0.4)
-0.4 (-0.8, -0.1)
-3.9 (-12.5, 4.7)
1.2 (0.0, 2.5)
4.1 (2.5, 5.7)
1.7 (-0.2, 3.7)
1.2 (-0.1, 2.5)
6.4 (2.2, 10.5)
8.2 (1.2, 15.2)
-0.1 (-4.4, 4.1)
-0.6 (-5.3, 4.2)
-1.0 (-4.2, 2.2)
3.7 (-2.1, 9.6)
-0.33 (-4.5, 3.8)
-1.4 (-4.5, 1.6)
P
0.78
1.00
0.67
0.19
0.10
<0.01
<0.01
<0.01
<0.01
0.37
0.05
<0.01
0.09
0.08
<0.01
0.02
0.95
0.81
0.53
0.21
0.88
0.35
Effect
size (d)
-0.05
0.00
0.08
-0.24
-0.30
-0.87
-0.81
-0.70
-0.51
-0.16
0.36
0.91
0.31
0.32
0.55
0.42
-0.01
-0.05
-0.11
0.23
-0.03
-0.17
BMI=Body Mass Index; BMI=Body Mass Index; ECBI=Eyberg Child Behavior Inventory (27); BPFAS=Behavioral Pediatrics Feeding Assessment Scale
(22); PPSFQ=Parent Perceived Feeding Stress Questionnaire (30); PSI=Parenting Stress Index (28, 29)
Effect size: d≥0.2=small; d≥0.5=medium; d≥0.8=large
Copyright © ESPGHAN and NASPGHAN. All rights reserved.