Promoting Routines of Exploration and Play during Mealtime: Estimated Effects and Identified Barriers

Abstract

Promoting Routines of Exploration and Play during Mealtime (Mealtime PREP) is an intervention designed to support healthy dietary variety in children. To estimate the effects of this intervention, we recruited 20 parents and children (aged 1–5 years) with sensory food aversions to participate in a pilot study. Parents were coached to enhance daily child meals using Mealtime PREP. Our primary outcome was acceptance of targeted food (number of bites) over time. Descriptive statistics and effect sizes are reported. Moderate effects were observed for acceptance of targeted food. Mealtime PREP warrants additional research to examine effects in larger, more diverse samples.

Early intervention to build healthy eating habits during the first five years of life is recommended as a critical component of early health promotion efforts (World Health Organization, 2020). Selective, or picky, eating is a common behavior pattern of early childhood, and feeding issues are a common reason for referral to pediatric occupational therapy services (Fisher, 2017). Young children with a limited food repertoire are at greater risk for psychopathological symptoms such as anxiety, depression and attention deficits (Zucker et al., 2015). Furthermore, food preferences developed in early childhood are likely to persist (Mascola et al., 2010). Consequently, practical interventions to build healthy dietary variety during early childhood is paramount to healthy habit development.

Behavioral modification techniques, including positive reinforcement and repeated exposure, improve acceptance of specific foods among young children (Cooke, 2007; Cruwys et al., 2015; Williams et al., 2010). Positive reinforcement (Skinner, 1938) is a strategy to increase desired behaviors, such as trying novel foods, by ensuring these behaviors are followed by positive consequences (Cooke et al., 2011; Remington et al., 2012). Repeated exposure to a specific food also improves child acceptance of that food (Remington et al., 2012; Wardle et al., 2003). However, to facilitate daily behavior change, it is critical that interventions empower caregivers to embed these techniques into existing routines. Evidence to support interventions that increase healthy habits and dietary variety by altering everyday routines in the home is limited (Marshall et al., 2015).

Frequent family meals are associated with higher fruit and vegetable intake, better family connectedness and lower rates of obesity in children (Berge et al., 2015; Brown et al., 2019; Caldwell et al., 2018a; Christian et al., 2013). Embedding positive reinforcement and repeated exposure to food within a consistent family meal routine could lead to improved dietary variety and decreased risk of nutritional problems, such as obesity and malnutrition, for young children. Providing opportunities to support acceptance of healthy foods and building mealtime routines in young children is a critical role for parents. Parents often recognize mealtime dysfunction, but face barriers (e.g. busy schedules, lack of support) to intervening within the natural context of the home (Golley et al., 2011). Consequently, approaches that embed evidence-based strategies into established routines in the home are needed to support improved healthy food acceptance and overall dietary variety daily without undue burden on the family.

Young children with sensory food aversions (SFA) are an ideal population for examining the effects of an intervention to improve healthy dietary variety, because they often avoid healthy foods based on sensory characteristics (Yang, 2017). Specifically, they are likely to refuse fruits and vegetables (Chatoor, 2009), have strong taste preferences, and are reluctant to try novel foods. Children with SFA: (1) refuse to eat specific foods with specific tastes, textures, smells, or appearances; (2) refuse upon the introduction of a different type of food; (3) eat better and more when offered preferred foods; and (4) have specific nutritional deficiencies or oral motor delay (Levine et al., 2011; Yang, 2017). Because children with SFA meet the criteria of persistent picky eaters, they are unlikely to try new foods without intervention (Toyama & Agras, 2016). Therefore, the risk of maturation bias, or spontaneous recovery from selective eating, is slightly attenuated within this population.

Although evidence on the effects of interventions for SFA is scarce, combining systematic and stepwise approaches with positive reinforcement is recommended (Yang, 2017). Food chaining, an approach that aims to expand a child’s food repertoire by emphasizing similarities between targeted and accepted food, has been found to improve dietary variety for children with food aversions. (Fishbein et al., 2006). However, this intervention does not emphasize the importance of shifting mealtime routines to promote long-term behavior change. We designed the Promoting Routines for Exploration and Play during Mealtime (Mealtime PREP) intervention using the transactional Person Environment Occupation (PEO) model to enhance mealtime at the level of the child (or family), the surrounding environment and structure, and the activity, or occupation, of eating. Mealtime PREP also applies powerful behavioral change strategies to mealtime routines to promote lasting behavior change (Caldwell et al., 2018b).

In this pilot study, we aimed to inform intervention optimization by generating effect sizes and identifying barriers in intervention pathways. We described targeted food acceptance before, during, and after parents were coached to enhance mealtimes of young children with SFA using the Promoting Routines of Exploration and Play during Mealtime (Mealtime PREP) intervention. Additionally, we examined risk of nutritional problems among the sample over time. Finally, we identified barriers to intervention application and assessment by examining data on recruitment, retention, data collection and treatment acceptance.

Methods

Participants

Families were recruited through local early intervention agencies, primary care physician referral, and social media advertisements. To be included, child participants had to be between the ages of 18 and 60 months (1–5 years) and meet all four diagnostic criteria of SFA: (1) consistently refused to eat certain foods with specific tastes, textures, temperatures or smells; (2) onset of food refusal occurred during the introduction of a new food; (3) eats without difficulty when offered preferred foods; and (4) presence of nutritional deficiency or oral motor delay (Zero to Three, 2005). We used the Diagnostic Classification of Mental Health and Developmental Disorders of Infancy and Early Childhood: Revised Edition (Zero to Three, 2005) to specify inclusion criteria for all children in our sample (even those older than three years) because even though SFA is often diagnosed prior to the age of three, it is observed in preschoolers and these criteria have been applied broadly to young children (Chatoor, 2009; Kerzner et al., 2015).

An occupational therapist screened each child participant to determine whether they met the clinical criteria for SFA using a parent interview, structured mealtime observation (Marcus & Breton, 2013) and the Nutrition Screening Tool for every Preschooler (NutriSTEP or NutriSTEP Toddler version; Simpson et al., 2008). Risk of nutritional deficiency was indicated by a score ≥20 on the nutrition screen. Oral motor delay was indicated through demonstration of immature or atypical oral motor skills (e.g., loss of food from inadequate lip closure, lack of lateral, diagonal, or rotary chewing patterns) based on structured observations. Inclusion criteria for parent participants included being at least 18 years old and having the ability speak and read English at a 6^th grade level. There were no exclusion criteria for this study.

Design

This study used a phased repeated measures design to collect rich daily data on the acceptance of targeted food in the home over time. All data collection and intervention delivery were completed in the home environment. We used a three phase (A-B-B¹) design to collect data on child targeted food acceptance before (A), during (B), and after (B¹) parents were coached to deliver the Mealtime PREP intervention. During each phase, parents were instructed to video record ten child meals; first, under normal circumstances over 2 weeks (A), then during the time they were being coached to deliver the Mealtime PREP intervention by an occupational therapist (within a 6-week period; B) and finally as they delivered the intervention without therapist support during a 2-week follow-up (B¹). Video-recorded meals were not time restricted, as parents were instructed to video record the entire mealtime. In the follow-up period, parents were encouraged to continue to practice all skills learned during the intervention phase.

Four occupational therapy clinicians with pediatric clinical experience ranging from one to thirteen years (mean = 7.75 years) led four to six parent-coaching sessions in the child’s home within a six week period. Intervention duration and frequency was modeled after local service delivery patterns to promote the feasibility of translation into current clinical practice. Each clinician completed at least eight hours of training on intervention delivery prior to working with families. Data on risk of nutritional problems were collected pre- and post-intervention delivery and intervention acceptability was rated post-intervention. This study underwent a formal ethics review and was approved by the university Institutional Review Board. All participants completed informed consent, with at least one parent providing written consent for participation (self and child). This trial is registered with ClinicalTrials.gov, NCT03138551.

Intervention

Mealtime PREP is a two-pronged intervention that focuses on both the parent(s) and child. Because parents are responsible for providing food options, organizing the environment, and scheduling child meals, they are a critical contributor to every child mealtime experience. Therefore, we developed an intervention in which occupational therapy clinicians coached parents to deliver evidence-based techniques during daily family meals.

Parents were coached to deliver each intervention component during mealtimes using a stepwise, behavioral activation approach. Behavioral activation is an effective method of promoting behavior change and establishing new routines (Cuijpers et al., 2007). The parent-coaching prong of the Mealtime PREP intervention incorporated four active ingredients of behavioral activation (1. skills training; 2. goal setting; 3. activity scheduling; and 4. activity monitoring) to help parents build a family meal routine enriched with techniques to promote child food acceptance (See exemplar in Table 1). Implementation of the Mealtime PREP intervention unfolds as the evidence-based strategies of positive reinforcement (e.g. praise, clapping, high fives; Green et al., 2015), repeated exposure to targeted foods (Cooke, 2007; Remington et al., 2012) and play (e.g., using broccoli as a paint brush in ranch dressing; “fishing” for targeted food with a fork; Coulthard & Sealy, 2017) are embedded into a consistent family meal routine incrementally.

Table 1.

Behavioral activation approach to parent-training: exemplar of therapist-led session.

Skills Training	During this session, Violet’s parents are coached on a new skill, Positive Reinforcement. During mealtime practice, Violet throws a green bean onto the floor. She is redirected, first by the occupational therapist, to put the green bean in a scrap bowl if she does not want it on her plate. Later in the meal, Violet throws a piece of chicken on the floor and her mother uses this strategy to redirect Violet to place the chicken in the scrap bowl if she no longer wants it on her plate.
Goal Setting	A new goal or “plan” is developed. Violet’s parents will positively reinforce 5 appropriate behaviors (and redirect inappropriate behaviors as needed) during 4 practice meals over the next week. This is added to their goal from last week to, “practice all 4 steps of their mealtime routine during 4 practice meals over the next week.”
Activity Scheduling	The occupational therapist works with Violet’s parents to identify four meals over the next week when they are going to practice these skills and write down this plan (including dates and approximate times) in their workbook.
Activity Monitoring	Each time Violet’s parents lead a practice meal, they log it in their workbook and video record the meal.

During each intervention session one of these strategies to promote child food acceptance was introduced and parents were provided with the opportunity to practice during a meal, role play, and troubleshoot issues. Therapist feedback included knowledge of performance (specific feedback provided while the parent practiced skills) and/or results (more general feedback provided after the meal or based on parent report of meals). Parents continued to practice skills by delivering intervention components in the home in between parent-coaching sessions. These practice meals were video recorded for data analysis, regardless of presence of the therapist.

The family meal routine was developed during the first intervention session and included a predictable beginning, middle, and end to each meal. Positive reinforcement was used to reward both behaviors categorized as “appropriate” by each family and acceptable alternate behaviors requested of the child when “inappropriate” behavior was observed. Child behaviors commonly labeled as inappropriate by parents included throwing and spitting food and standing up or leaving the table. Behaviors used frequently as acceptable alternatives included placing food in a scrap bowl, signing or stating, “all done,” or taking a drink. Parents also modeled various ways to interact (Toomey, 2010) with food and encouraged their child to explore and/or play with food.

Parents were provided with a copy of the mealtime routine and a list of ways to interact with food to facilitate intervention adherence during meals. The intervention was customized to accommodate a variety of mealtime environments, address unique child behaviors and incorporate play that was meaningful to each child (See Exemplar in Table 2). Parents were instructed to avoid use of screens (i.e., phones, televisions, tablets) during each intervention meal. The Mealtime PREP intervention has demonstrated acceptable feasibility in terms of parental treatment acceptability and intervention adherence (Caldwell et al., 2018b).

Table 2.

Exemplar of parent-mediated intervention components.

Family Meal Routines	Violet is given a warning, “We are going to get ready for dinner in two minutes.” After two minutes, Violet washes her hands and climbs into her highchair. She helps to self-serve macaroni and cheese (preferred) and chicken (targeted) onto her plate. Her mother and father also self-serve a portion of each food onto their plates. When Violet is finished eating and playing, she says, “all-done,” and helps with clean-up.
Positive Reinforcement	During the meal, Violet screams when the sauce from the macaroni and cheese touches her elbow. Her mother redirects her to use a napkin to wipe the sauce away. She wipes the sauce from her arm and her mother praises her. Violet also tries a bite of chicken and receives a high five from her father.
Food Exploration and Play	Throughout the meal, Violet’s mother and father model interacting with food in different ways (making a smiley face with green beans, pretending bites of chicken are “swimming” in the macaroni), and praise each other for trying and interacting with all foods. Violet is invited to join in, and “fishes” some chicken out of her macaroni and cheese using her fork. She then decides to try a bite!

Measures

Targeted food acceptance was assessed through observation of video-recorded meals as the number of bites of targeted food consumed. “Bite” included any event when the child accepted food into his/her mouth, and it did not come back out of the oral cavity. The parent and therapist identified “targeted foods” collaboratively prior to initiating intervention as healthy foods the child refused to try or spit out after one bite. To promote cultural relevance, parents were encouraged to select healthy foods that are offered frequently in their home. All video-recorded child meals were coded by a team of five graduate student research assistants and the principal investigator for number of bites of targeted foods. Inter-rater reliability of the coding system was found to be excellent, ICC = .948; 95% CI = .894–.974 (Caldwell, 2017). For each phase, we examined the number of bites accepted during each meal and calculated an average number of bites of targeted food per meal to assess change over time in this variable.

Risk of nutritional problems was measured through parent completion of the NutriSTEP or NutriSTEP Toddler (Simpson et al., 2008), based on the age of child, pre- and post-intervention. This tool assesses nutrition risk by accounting for dietary variety, behavioral risk factors (e.g., distraction during meals), and parent perception of growth. The NutriSTEP is a reliable and valid screen of nutrition risk in preschoolers and toddlers (Simpson et al., 2008). The toddler version is used for children between the ages of 18 and 36 months and the preschool version used for children aged 36–60 months. Cutoff scores have been established to determine risk of nutritional problems (low < 20, moderate 20–24, and high ≥25).

Intervention acceptability was measured using the Treatment Acceptability Questionnaire (TAQ; Krain et al., 2005). The TAQ is a brief, eight-item assessment of treatment acceptability that has been validated for use with parents of young children (Krain et al., 2005). For each intervention acceptability statement, respondents rate how much they agree on a scale of 1 (strongly disagree) to 6 (strongly agree) as it relates to their child (e.g., “This treatment should be effective at changing my child’s behavior,” “I like this treatment”). Higher scores represent greater intervention acceptability, with a score greater than 28 suggested to indicate that an intervention demonstrates acceptability (Tarnowski & Simonian, 1992). We also collected demographic information (parental education, parental employment status, parental marital status, household income, child age, child race, and child gender) at baseline using a paper form. Parents were asked to report medical issues, diagnoses, and other feeding issues (e.g., gastroesophageal reflux, intolerances) on this form, as formal testing for other feeding issues was not completed.

Data analysis

We used the following Medical Research Council intervention optimization principles to develop our data analysis plan: (1) Generating (tentative) estimates of effect size and (2) Identifying barriers or rate limiting steps in intervention pathways (Campbell et al., 2007).

Generating estimates of effect size

Descriptive statistics and effect sizes (Cohen’s d) were used to describe within-subject changes in targeted food acceptance (average bites per meal during pre- and post-intervention phases) and risk of nutritional problems over time. An alternate formula was used to estimate effect sizes that considers the correlation between pre- and post-intervention scores that is most appropriate for calculating single-group pretest-posttest effects (Lenhard & Lenhard, 2016; Morris, 2008; Morris & DeShon, 2002). We also calculated the tentative effects of the intervention based on age categorization as toddlers (18–36 months) and preschoolers (36–60 months). Descriptive statistical analyses were completed in IBM SPSS Statistics for Windows, version 24 (IBM Corp, 2016).

Identifying barriers to application of intervention

Recruitment, retention rate, and success of planned data collection were examined as a first step to identifying potential barriers to intervention application and evaluation. Additionally, we looked at responses to each question in the TAQ to identify areas in which participants were less satisfied with the intervention. We examined any scores that did not reflect agreement (<4) to identify potential barriers.

Results

Twenty-two of 25 families who were screened were eligible for participation in our study. Two families withdrew from the study prior to intervention based on personal (N= 1) and medical (N= 1) issues. Twenty families completed the Mealtime PREP intervention and are included in our analyses. Child participants ranged in age from 19 to 54 months. Of the 20 children in our sample, 14 demonstrated elevated risk of nutritional problems and 6 met inclusion criteria based on oral motor delay (Table 3).

Table 3.

Sociodemographic characteristics of intervention completers (N = 20).

Characteristic	n (%)
Child Gender
Male	15 (75)
Child Age
18–35 months	15 (75)
36–60 months	5 (25)
Child Race/Ethnic group	19 (95)
Non-Hispanic white
Child Diagnosis
Autism	1 (5)
Developmental delay	4 (20)
Gastrointestinal issues	6 (30)
None reported	9 (45)
Household Income ($)	3 (15)
50,000–75,000
75,001–100,000	5 (25)
>100,000	12 (60)
Highest Parent Education (n = 36)	3 (8)
Vocational degree
Bachelor’s degree	19 (53)
Graduate degree	14 (39)
Parent Employment Status (n = 36)
Full-time	27 (75)
Part-time	3 (8)
Not employed	5 (14)
Retired	1 (3)

Generating estimates of effect size

Targeted food acceptance

On average, targeted food lists included 11 foods (SD = 4.2). Most healthy targeted foods were proteins (34%), vegetables (21%), or fruits (15%). Average targeted bites per meal increased over time. During the pre-intervention phase, children accepted, on average, 1.53 bites of targeted food per meal, with 13 children accepting zero bites of targeted food throughout this entire phase. This average increased to 2.81 bites per meal during the intervention period and 3.48 bites of targeted food per meal during the post-intervention period. Thirteen of 20 children improved targeted food acceptance, indicating that they were willing to try foods previously refused. We observed a moderate effect (d =.52) for increasing the number of bites of targeted food accepted from baseline to post-intervention across our sample. It is notable, that when tentative effect sizes were calculated by age cohort, toddlers demonstrated a large effect (d = 0.86) for increasing bites of targeted food per meal from pre-intervention (M = 1.06) to post-intervention (M = 3.84) while preschoolers remained relatively stable (d = 0.02) in acceptance of targeted food from pre- (M = 2.93) to post- (M = 2.37) intervention (Figure 1).

Figure 1.

Child bites of targeted foods per meal over time.

Risk of nutritional problems

We observed a difference between pre-intervention (M = 21.75, SD = 5.47) and post-intervention (M = 18.15, SD = 4.10) scores on the NutriSTEP assessment; d = 0.76. This indicates a moderate effect for reducing risk of nutritional problems among the children in our sample. When effect sizes were calculated by age cohort (18–36 months; 36–60 months), we found similar effects (d = 0.53; d = 0.58). It is clinically noteworthy that risk categorization shifted over time with six children classified as low risk, 10 categorized as moderate risk, and four categorized as high risk for nutritional problems pre-intervention. These numbers improved to 13 in the low-risk group, six in the moderate-risk group, and one in the high-risk group after intervention.

Identifying barriers to application of intervention

Twenty of the 22 eligible families (91%) completed the study. Of these 20 families, we were able to collect data on targeted food acceptance during mealtimes in the home for all but two children during all three phases (90%; before, during, and after intervention). For both children, missing data on targeted food acceptance occurred after intervention (one due to video camera malfunction and the other due to conflicting parental demands). In general, recruitment, retention, and ability to collect data on outcomes of interest were not significant barriers to the application of the intervention. However, the diversity of our sample, in terms of race, ethnicity and income-status, was limited. Lack of diversity in the sample was noted as a potential barrier to the application of the intervention.

All 20 parents who completed the study rated the Mealtime PREP intervention as acceptable (M = 43.6, SD = 4.39). Only two parents provided ratings lower than four for any of the eight items on the scale. One parent provided poor acceptability ratings regarding acceptability of the intervention for their child’s behavior (2), the effectiveness of the treatment to change behavior (3), and whether this treatment was a good way to handle their child’s problem (2). Of note, this child was 19 months old, being evaluated by early intervention services for an oral motor and speech delay and did not show improvement for targeted food acceptance nor risk of nutrition-related problems. The other parent who gave low ratings only provided a less than acceptable rating for one item related to whether their child’s behavior was troublesome enough to justify use of the treatment (2). This child was 25 months old, did demonstrate improvement in targeted food acceptance, but did not show lessened risk of nutritional problems (demonstrated minimal risk both pre- and post-intervention). Both parents were highly educated, married, and reported household income >$75,000 annually.

Discussion

Improving dietary variety for young children with SFA is challenging, however, in a small sample of young children with SFA, the Mealtime PREP intervention yielded improvements in targeted food acceptance and decreases in risk of nutritional problems. These observed moderate effects are meaningful because young children with SFA have restricted diets and are often unwilling to explore new foods. Notably, healthy children have been observed to consume approximately 2–3 bites of a novel vegetable upon initial exposure, on average (Fildes et al., 2014). This indicates that our sample was more resistant to trying new food, with 13 of 20 children accepting zero bites of targeted foods throughout the entire pre-intervention period. Despite this resistance, we observed an average improvement of two more bites of targeted food per meal, which is comparable to similar studies in children without sensory food aversions (range 1–6; Fildes et al., 2014; Nederkoorn et al., 2018; Wardle et al., 2003). It is important to acknowledge that Mealtime PREP was not a perfect fit for all families; therefore, we plan to use lessons learned from this pilot study to optimize this intervention prior to formal hypotheses testing in larger, more diverse, sample.

Lesson #1: We learned that Mealtime PREP may be more effective at changing behavior if delivered to families prior to the age of three. Toddlers demonstrated large effects for acceptance of targeted foods during meals over time, while we did not observe much change among preschoolers. Although our sample of preschoolers was quite small, these findings are consistent with previous work that shows that a child’s willingness to try and gain preferences for new foods decreases over time, (Mascola et al., 2010) and may signal that the Mealtime PREP intervention requires modification to best meet the needs of children between the ages of three and five years. Because these children have more firmly established preferences, they may require more time to benefit from repeated exposure and exploration of healthy foods in the home (Ding et al., 2012). Additionally, enhancing the current intervention with more opportunities for preschool-age children to participate in meal preparation may improve their acceptance of novel, targeted foods (Van der Horst et al., 2014).

Lesson #2: We learned that the Mealtime PREP intervention was widely rated as acceptable by parents in our sample, in congruence with prior reports (Caldwell et al., 2018b). However, this program may not be a perfect fit for all participants. Interestingly, our data signals that some children with SFA may not be a good fit for Mealtime PREP because their behavior does not justify the use of this treatment method, while others may have additional needs that are not being met by our approach. To recruit children who are a good match for the Mealtime PREP intervention, we must reflect on the focus of this intervention. Mealtime PREP promotes improved food acceptance, and therefore dietary variety, by focusing on child behavior, exploration, and play. It does not, however, focus on improving the oral motor or cognitive skills of child participants. Therefore, to meet the needs of all children with SFA, many of whom face additional challenges due to developmental delays, Mealtime PREP will need to be modified and/or supplemented with additional intervention strategies. For example, children with ASD may be resistant to treatment and frequently require an approach that incorporates hunger induction and nutritional supplementation in addition to sensory exploration and desensitization (Kerzner et al., 2015). In the current form, Mealtime PREP may be best suited to meet the needs of young children demonstrating SFA without the co-occurrence of developmental delays. Future trials should strive to examine the needs of children with diagnoses that commonly demonstrate SFA, such as ASD (Johnson et al., 2014) and/or Down syndrome (Oliveira et al., 2010).

Lesson # 3. Recruitment methods used to identify potential participants (i.e., ads posted on social media and flyers posted in the community) may have limited the reach of the Mealtime PREP intervention. This method yielded a largely homogeneous sample in terms of race, ethnicity, parental education, and household income. Due to this lack of diversity, it is important to consider how these factors might impact the observed preliminary effects for targeted food acceptance and risk of nutritional problems. This is a notable limitation of this pilot study, as children from low-income homes have considerably greater risk of nutrition-related problems (Rogers et al., 2015). Lack of knowledge about clinical studies has been identified as a barrier to participation among racial and ethnic minorities (Byrne et al., 2014); therefore, we plan to collaborate with community partners, such as the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) and local community centers in diverse neighborhoods to recruit participants for future studies. More data are needed to optimize the Mealtime PREP intervention to meet the needs of a diverse population of children.

Limitations of this study include a small sample size and lack of diversity (in terms of demographics and diagnoses), which limits our ability to make inferences about the effects of Mealtime PREP to the larger population of children with SFA. Additionally, we did not formally assess children in our sample for other childhood eating disorders and relied on parent report of additional feeding issues, development delays or cognitive deficits, which may have influenced response to intervention. Clinical experience of the therapist varied and is another potential confounding factor that will be examined in future trials with larger samples. Lastly, we did not control for the potential influence of therapist presence (or absence) during meals or the variation of length of meals during the intervention phase when coding acceptance of targeted bites.

Conclusion

In summary, we learned that the Mealtime PREP intervention is a promising treatment approach to improve targeted food acceptance and decrease risk of nutritional problems for children with sensory food aversions. It is recommended that clinicians working with young children with SFA include repeated exposure to targeted foods, positive reinforcement, and sensory-based play in their treatment plans to improve dietary variety (Cooke, 2007; Coulthard & Sealy, 2017; Remington et al., 2012). We have learned valuable lessons that will drive future decisions related to recruitment strategy, intervention optimization, and adaptation of our approach to meet the needs of children with developmental delays and those who are older than three years. Additional research is required to determine the efficacy of the Mealtime PREP intervention. Our future trials will strive to recruit a more diverse sample, in terms of race, ethnicity and socioeconomic status, to generate estimates of effect and identify additional barriers to intervention application, as well as resources required to overcome these barrier

About the authors

Angela Caldwell is an occupational therapist and assistant professor at the University of Pittsburgh. Her primary area of expertise is pediatrics with an emphasis on health promotion during early childhood. Caldwell is particularly interested in the development and implementation of interventions to reduce childhood health disparities. Co-authors Skidmore, Terhorst, Raina, Rogers, Danford, and Bendixen contributed as mentors with expertise in intervention development, research methodology, health promotion, and pediatrics.

Data availability statement

The associated data set is available upon request and approval of the corresponding author pending an approved data use agreement.