A Scoping Review of Tools to Assess Diet in Children and Adolescents with Autism Spectrum Disorder

Eating is considered one of the activities of daily living most affected by autism spectrum disorder (ASD) in children and adolescents and, therefore, needs to be thoroughly assessed using specific tools. The aim of this scoping review was to describe the most widely used tool to assess diet in children and adolescents with ASD. A search was conducted on PubMed, Scopus, EMBASE, Web of Science and PsycINFO databases. Two authors screened the articles and included all randomized or non-randomized studies published in English or Spanish in the last five years in which the diet of children and adolescents with ASD was assessed. Fifteen studies were included in this review. Mealtime behaviour was the most assessed variable in the included studies (n = 7). Thirteen different assessment tools were identified to evaluate the diet of children and adolescents with ASD, mainly at ages 2–12 (n = 11). The Brief Assessment scale for Mealtime Behavior in Children (BAMBI) and 24-h recalls were the most commonly used dietary assessment tools in the included studies. Our results can help professionals in the selection of an optimal scale to assess diet in children and adolescents with ASD.


Introduction
Autism spectrum disorder (ASD) is defined as a neurodevelopmental disorder mainly characterized by persistent deficits in social communication, difficulties in social interaction and a predisposition towards a pattern of stereotyped behaviours, interests and activities [1]. The etiological factors are diverse, and include epigenetic and environmental agents, although the definitive cause of ASD remains unknown [2,3]. These disorders manifest themselves at an early age, but are usually diagnosed later in childhood, at approximately 4 years of age [4]. According to the 2023 Centres for Disease Control and Prevention (CDC) report, the United States prevalence of ASD is 1 in 36 children, it is higher in boys than in girls and is continuously increasing [5,6].
Eating is one of the most affected daily life activities among children and adolescents with ASD, mainly because they tend to present hypersensitivity to some food textures, colours or tastes, which prevent them from trying new and unfamiliar foods [7,8]. This sensory alteration is usually manifested as repetitive, problematic and challenging behaviours at mealtimes, which may result in a restricted diet in terms of number and variety of foods [9][10][11]. This unvaried diet can even lead to nutrient inadequacy [12,13] and overweight [14,15], both of which can have negative effects on children's and adolescents' health [16]. Thus, there is an urgent need to develop dietary assessment tools for children and adolescents with ASD, which can be completed by parents and caregivers, to achieve a more accurate intervention in this issue.
There is evidence which supports that existing standardised methods for assessing feeding problems, eating behaviours and diet quality in ASD are very limited [17,18]. As dietary patterns are established in childhood and continue into adulthood [19], it is important to know and to use valid assessment tools to evaluate diet at an early age to identify the main nutrition necessities and to promote an optimal intervention. Currently, there are some published reviews on dietary assessment tools in children and adolescents with ASD. De Souza Silva et al. [20] carried out a Systematic Review (SR) aimed at evaluating the methodological quality of dietary assessment methods used to evaluate the diet intake of children and adolescents with ASD in clinical and epidemiological studies. This SR included eighty-nine articles and the results showed that most of them had a low-quality score. Holloway et al. [21] carried out a Scoping Review aimed at examining the evidence of validity and availability of measurement tools that evaluate usual dietary intakes and physical activity behaviours among individuals with ASD. In this review, one hundred and thirteen articles were included, and the results showed the need for more validated dietary tools in the ASD population. In this sense, we seek to perform a Scoping Review to complement the existent scientific evidence by answering the following research question: Which tool has been the most used to assess the diet of children and adolescents with ASD in intervention studies published in the last five years? The objective of this Scoping Review is to provide physicians and researchers with an updated source of information on the most widely used dietary assessment tool for children and adolescents with ASD.

Materials and Methods
We carried out a Scoping Review following the standards of the Cochrane Handbooks Version 6.2, 2021 [22] and the recommendations of the PRISMA Extension for Scoping Reviews (PRISMA-ScR) [23]. We conducted this type of review because our research question is broad and therefore could not have been addressed by a Systematic Review (SR). SRs are used to answer specific research questions which are usually related to the effectiveness, costs or effects of particular interventions [24,25]. We have not published a protocol of this review.

Search Strategy
On 10 August 2023, two of the authors of this review conducted a literature search in five databases: PubMed, Scopus, Web of Science, EMBASE and PsycINFO. We used the same search strategy and the same search terms in all these databases, using four different combinations: (1) (ASD OR autism OR autistic OR autistic OR asperger OR rett OR pervasive OR disintegrative), (2) (food OR diet), (3) (1 AND 2), (4) 1 AND 2 in the last 5 years. Time filtering (last 5 years) was used for all searches (Table 1). We applied this time filter because we wanted to provide a current synthesis of information. The time frame for considering an article to be current or not varies between disciplines, although in health and medical sciences it is recommended to use references from the last 5 years [26].

Review Criteria and Study Selection
We established for this review the following inclusion criteria: (1) articles published in English or Spanish; (2) articles available in full text; (3) articles whose study population consisted of persons ≤18 years old with Autism, Rett Syndrome, Asperger Syndrome, Disintegrative Disorder or Pervasive Developmental Disorder; (4) articles in which diet was evaluated; (5) articles with randomised or non-randomised study design. It should be noted that each intervention study has only been included once. In other words, publications derived from the same original intervention study were not included.
The process of study collection and subsequent data extraction was carried out independently by two authors (COB and LMCG), with a third author (LTC) intervening in case of discordance. In a Microsoft Excel spreadsheet, we downloaded all article titles retrieved from the different databases. On the Excel database, all studies were examined and selected for this scoping review by two authors using a four-stage screening process: elimination of duplicate articles, screening by title, screening by abstract and screening by full text. We then created the PRISMA Flow Diagram using the free-to-use web-based online tool available on the PRISMA website [27].

Data Extraction
In order to facilitate data extraction and avoid subjectivity on the part of the authors, we designed the items included in the tables a priori in accordance with the Cochrane Handbooks recommendations [22]. In one of the tables, we listed the general characteristics of the included studies as follows: author, year of publication, study design, study sample, country, participants, intervention/comparator, evaluation and dietary study outcomes. In another table, we listed the characteristics of the assessment tools used in the included studies as follows: dietary assessment tool used, author, year of publication, study participants, dietary assessment tool description, tool scores and assessment manager. Finally, we included a table containing items related to the risk of bias in the included studies, such as main limitations, funding sources and conflicts of interest.

Quality Assessment
We did not assess the quality of the studies included in this review, as it is not a mandatory requirement in scoping reviews [28]. Nevertheless, we have provided a table in which we have extracted and synthesised information related to the quality of the studies. In addition, we have described the main limitations of the included studies in the results section of this scoping review in considerable depth to allow readers to assess the results of the review in a more critical way.

Results
The initial search strategy retrieved a total of 13,374 articles published in the last 5 years. After the removal of duplicate articles, 5806 articles remained for screening. During the screening, 4651 articles were discarded by title, 1083 articles by abstract and 54 articles by full text. In the full-text screening, three articles [29][30][31] were excluded as they were publications derived from previous intervention studies which were included [32,33]. At the end of this process 15 articles were included in this scoping review ( Figure 1).

3-day food records
Piwowarczyk et al. [44], 2019 66 children with ASD These records were obtained in two different moments of follow-up (week 2-4 and at week 12).
Adherence to the gluten-containing diet was described as the consumption in more than one meal every day of some gluten-containing foods. Adherence to the gluten-free diet was described as suitable when no intake of gluten was stated in the food record.

Study coordinators and psychologist
Javadfar et al. [37], 2020 43 children with ASD These records were collected at baseline, 8 and 15 weeks of the intervention.
Nutritionist IV software evaluated dietary intake. Energy and micro/ macronutrients were calculated. Higher scores reflect higher intakes. Five-point Likert scale with 35 items, from "never" to "always". Higher levels of fussy eating were indicated by a higher score for food fussiness. High levels of enjoyment were indicated by a high score for enjoyment of food.

The BAMBI and the BAMBI-R
One of the most commonly used dietary assessment tool in the included studies was the BAMBI and its revised version the BAMBI-R (n = 4) [33,35,43,45] (Table 3). In the article by Galpin et al. [43], this tool was used to assess food selectivity in children and adolescents with ASD. Patton et al. [35] used the BAMBI to assess eating problems and eating behaviour in relation with the severity of ASD symptoms. Finally, Chung et al. [45] used this tool to assess how children and adolescents' eating behaviour, diet and sensitivity may vary depending on the physical appearance of food.
In all these articles, the BAMBI was interpreted using the same scores, including a total score and three subscale scores. Questions can be answered with a Likert scale ranging from 1 (never/rarely) to 5 (at almost every meal) and a higher score represented a higher frequency of problematic behaviours and food selectivity. For its administration, health professionals asked the main caregivers to fill in the BAMBI based on observations during the mealtimes of children and adolescents with ASD.
The BAMBI-R is the revised version of the BAMBI, and their scores and interpretation are thus very similar. The main difference between these tools is that the BAMBI-R can be administered by teachers in the school setting. The BAMBI-R was less widely used (n = 1) [33] among included studies than the BAMBI (n = 3) [35,43,45], and it was used to evaluate the feasibility and effectiveness of a structured multidisciplinary intervention designed for children and adolescents with ASD with moderate food selectivity.

Twenty-four-h Dietary Recall
Another of the most commonly used dietary assessment tool in the included studies were 24-h dietary recalls (n = 4) [36,41,42,46] (Table 3). This dietary tool is a food diary that provides detailed information on portion sizes, preparation methods and quantities of food and beverages consumed by children in an entire day. In all the articles, the dietary recalls were filled out by parents. In the articles by Kim et al. [46], de la Torre-Aguilar et al. [42] and Kral et al. [36], three 24-h dietary recalls for two weekdays and a weekend day were used. In contrast, in the article by González-Domenech et al. [41] two 24-h dietary recalls per week were used. The type of analysis of 24-h dietary recalls was only specified in two articles. In the article by Kim et al. [46], the results were analysed using CAN-PRO 4.0, whereas in the article by Kral et al. [36], the University of Minnesota Nutrition Coordinating Center's Food and Nutrient Database was used.
Limitations in relation to the type of questionnaires used or assessment performed were also reported in several articles. Some examples of this kind of limitations are the non-use of standardised questionnaires or the lack of comprehensive assessment of children's selective eating and/or dietary variety (n = 3) [33,35,43], the use of questionnaires completed by parents (n = 1) [33], as well as the lack of direct observation by researchers (n = 1) [32].
Finally, limitations in relation to the intervention procedure were also reported, such as the participants' difficulty in following the recommended diet (n = 3) [39,41,44], failure to correctly weigh or record children and adolescents' food intake (n = 2) [32,45] or a short duration of the intervention (n = 1) [36] (Table 4). No financial support of any kind. None declared.
-The severity of CP and ASD diagnoses is unknown.
-A lack of categorization based on the severity of motor impairment.
Not stated. None declared.
-Low generalizability of the results due to the heterogeneous group.
-A lack of control group.
-A lack of baseline control period.
-A lack of meaningful standardized assessment measures.
Not stated. None declared.
-Convenient sampling method was used to select participants.
-Large variability among participants because the different characteristics included in the feeding problems.
-The observed variables did not meet normal distribution.
-Non-significant changes in nutritional intake.
Not stated. None declared.
-Interaction between food selectivity and disruptive behaviour.
-A lack of standardized measures to assess dietary variety.
-Changes in food selectivity severity were not assessed.
-Treatment assignment was not blinded from parents.
Eunice Kennedy Shriver National Institute of Child Health and Human Development supported the study by grants to Emory University (MH081148).
None declared. Yamane et al. [40], 2019 -The support group classification was only based on observations. Not stated. None declared.
-The supplementation period was short.
Vice chancellor of research and technology of Kermanshah University of Medical Sciences funded the study as a thesis proposal for the MSc degree.
None declared.
-Low generalizability of the snack preparation to all type of foods.
-A lack of control group.
-A lack of statistical significance.
-Only three fruits and three vegetables were studied.
No financial support of any kind. None declared.
Patton et al. [35], 2020 -Families knew they were being observed during meals, fact that can reduce mealtime interactions.
-Evaluators were not blinded.
-Low internal consistency on some of the BAMBI subscales.
-Specific questionnaires to assess sensory sensitivity were not used.

Discussion
This scoping review aimed to identify the tools that have been used in experimental studies over the last five years to assess the diet of children and adolescents with ASD. Among the thirteen studies included in this Scoping Review, thirteen different dietary assessment tools were used: BAMBI, BAMBI-R, DINE, SAPS, FFQ, CEBQ, CGI-I, CEBI, 3-day food records, 24-h dietary recalls, home eating records, food indices and food questionnaires. The most commonly used tools were the BAMBI/BAMBI-R which was used to assess mealtime behaviour and food selectivity in children and adolescents with ASD and 24-h dietary recalls which were used to assess the dietary intake.
Most of the studies included in this review were conducted in the United States (n = 4). In this country, 1 in 36 children and adolescents have ASD [5,6], and 17% of this population suffers from obesity [47]. This high prevalence represents a significant health problem which is of great scientific interest. In fact, the Journal of the American Dietetic Association reflects that concerns about the adequacy of the diet of children and adolescents with ASD and the management of dietary selectivity are the main reasons for referral to nutrition services in the United States [48], which may partially justify a greater number of studies involving a wider variety of dietary assessment tools. The United States has a wide range of educational, medical, behavioural, nutritional and social services to identify cases with ASD and understand their needs and those of their families [49], which in some way may facilitate research on dietetics and nutrition in ASD.
In general, the dietary assessment tools used in the included studies were used to assess the diet of children and adolescents with ASD as a global concept, including sensory or behavioural aspects. In fact, mealtime behaviour was the most frequently assessed variable in the included articles. In this sense, different authors indicate that feeding difficulties in children and adolescents with ASD are mainly related to sensory and/or behavioural problems [50,51]. These problems may be manifested during mealtimes as playing with food, eating very slowly, filling their mouths with food, closing their mouths tightly, swallowing food without chewing or gagging continuously [10,50]. The high prevalence of these behaviours in children and adolescents with ASD [51] and their relationship with feeding can justify that the majority of the dietary assessment tools used include not only nutritional items but also behavioural and sensory ones.
The dietary assessment tools most commonly used in the included studies were the BAMBI/BAMBI-R and 24-h dietary recalls. On the one hand, one of the main arguments put forward by authors for the frequent use of the BAMBI is that it allows for the assessment of several areas affected in the feeding of children with ASD, such as food selectivity and behavioural characteristics of ASD related to sensory responses to food variability [35,43,45]. This can also be explained by the fact that this tool has been described as a one that addresses limitations existing in other tools, such as not measuring behavioural aspects in ASD which, as mentioned above, are very important in feeding [52]. In addition, the BAMBI presents good internal consistency, high validity and high test-retest reliability as well as a clear and solid structure for the measurement of the behaviours of this population [53]. Finally, a fact that may influence the more frequent use of this tool is its relatively easy administration, as it is completed by the parents of children and adolescents with ASD [52].
On the other hand, we have found a high use of 24-h dietary recalls to assess dietary intake in children and adolescents with ASD among included studies. Dietary intake is a very complex health behaviour, with large daily variations in the foods and beverages a person consumes, which makes its assessment complex. Twenty-four-h dietary recalls have been widely used in epidemiological studies to assess dietary intake due to its validity, high response rate, and simplicity [54]. This type of dietary tool is usually self-completed, but in the case of the paediatric population with or without ASD, 24-h dietary recalls are completed by parents or caregivers. This is because children are not familiar with the different methods of food preparation and do not have fully developed writing skills [55]. In this sense, a recently published review concluded that, although 24-h dietary recalls can underestimate energy intake, data collected using self-reported dietary assessment methods in children are highly valuable [56]. It should be noted that the collection of dietary information can be even more important and valuable in children and adolescents with ASD, as their communication limitations and sensory processing difficulties may affect diet, which can result in inadequate nutrient intake [16].
This scoping review has some limitations that need to be mentioned when interpreting our results. These results could be influenced by limitations common to most reviews, such as the lack of information reported in the included studies, publication bias, which limits null results of the interventions, and selection bias. We have possibly increased selection bias by only including articles published in the last five years, with full-text available and written in English or Spanish. In recent years, the classification of ASD has changed. Currently, it is considered a neurodevelopmental disorder in the DSM-5, and the specific disorders that were included in the ASD are no longer used. However, some recently published articles continue using these specific ASD. Thus, we decided to use the terms included in the DSM-IV definition (Autism, Rett Syndrome, Asperger Syndrome, Disintegrative Disorder or Pervasive Developmental Disorder) in our search strategy in order to not overlook some potential articles for our review. Finally, with regard to included studies, we need to point out that we have only included randomized and non-randomized clinical trials, which could contain biases related to this type of study design.
Our review also has some strengths. As far as we know, it is the first review that aims to describe the tools that have been used the most to assess the diet of children and adolescents with ASD in experimental studies. It is an up-to-date source of information as we have focused on research from the last 5 years. This review could be very useful for professionals involved in ASD treatment to select and use the dietary assessment that better fits their intervention objectives, as it provides a clear synthesis of different dietary assessment tools. Furthermore, this study has also identified some gaps in knowledge: the need for more studies to be carried out in Spain and at a European level and the need for studies with a larger sample size and with a greater post-intervention follow-up.

Conclusions
Two dietary assessment tools were the most widely used in children and adolescents with ASD. On the one hand, the BAMBI was widely used, possibly because it not only assesses diet but also other behavioural aspects that can alter different activities of daily living in this population. On the other hand, 24-h recalls were widely used, possibly because they assess a complex behaviour, namely dietary intake, in a simple parent-reported way. Dietary assessment tools in this population tend to provide information on food selectivity and mealtime behaviours, not only of the children and adolescents but also of their families and the environment in which mealtime takes place. The results of this scoping review could help different professionals assess diet in a more exhaustive and comprehensive manner, as well as promote the development and validation of more dietary assessment tools for children and adolescents with ASD.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.