A Narrative Review of Dietary Assessment Tools for Preschool-Aged Children in the Home Environment

Preschool-aged children in the U.S. have suboptimal diets. Interventions to improve child nutrition focus on parents and their role in shaping social and physical home environments, which influence children’s eating behaviors. Dietary assessment tools selected to measure intervention objectives, and how results are interpreted in key findings, are essential when examining children’s diets. The objectives of this review were to (1) describe dietary assessment tools used in intervention studies in young children focused within the home environment; and (2) examine how the application of these dietary assessment tools addressed intervention objectives. PubMed and Web of Science were searched for English-language nutrition intervention studies that included children aged 2–5 years, had a home environment component, used a dietary assessment tool, and reported on diet-related outcomes. Seventeen studies were included. Intervention objectives focused on overall diet, specific food groups, eating occasions, and obesity prevention/treatment. Concordance of key findings with intervention objectives, type of tool used, and multiple tools within the same study varied with 8 studies aligning in objective and tool, 1 discordant in both, and 8 partially concordant or too broad to determine. This review highlights current challenges in measuring dietary intake in preschoolers and provides recommendations for alternative applications and strategies.


Introduction
Early childhood is a period for growth and development and when eating habits and dietary behaviors are formed. Nutrition is a critical contributor to overall health and plays a vital role in the prevention of diet-related chronic diseases, such as obesity [1]. Unfortunately, young children in the United States (U.S.) have poor diet quality [2] and they continue to fall short in achieving adequate intake of nutritious foods, particularly vegetables, whole grains, and dairy [3]. Conversely, children have high consumption of added sugars, sodium and saturated fat when compared to the recommendations of the Dietary Guidelines for Americans [3]. U.S. children aged 2-to 5 years consume approximately 75 percent of daily energy intake at home, emphasizing the important role parents and the home food environment have on influencing children's dietary intake [4,5]. The home is the child's first food and eating environment, is fundamental in shaping the emergence of eating habits in early childhood and continues to be a critical environment throughout childhood [5].
Interventions targeting dietary intake and eating behaviors of preschool-aged children are increasingly focused on parents as agents of change, largely due to the extensive evidence supporting parents' influence on shaping children's eating and growth [6][7][8]. In

FFQ/ FFQ Screener
Defined list of foods and beverages; asks frequency of consumption over an extended timeframe. Respondents choose from close-ended, multiple-choice options. Usually selfadministered but can be interviewer administered/assisted. FFQ contains a comprehensive list of items (~120-180 items). FFQ screeners contain an abbreviated list of the specified items (~20 items) and can be targeted to a specific food group or nutrient. Sample period: 1 W to 6 M (vs. 12 M for adults) Complex to navigate and literacy level of respondent should be considered. The FFQ Screener provides only high-level view of intake. List of food items may not include foods commonly consumed in some cultures or child-friendly items. *

Food Checklist
Defined list of foods and beverages for which respondents are asked to check which of the specified items were consumed over a specified time period. It may also ask about behavioral habits (e.g., reading nutrition labels). Portion sizes may be captured. Sample period: single or multiple days Low participant burden, although literacy level of respondent should be considered.
List of food items may not include foods commonly consumed in some cultures or child-friendly items.
Energy: calories/kilojoules; Macronutrients: proteins, carbohydrates, fats; Micronutrients: vitamins and minerals; Score/Index: diet diversity, Healthy Eating Index, etc. * Provides energy as an output but less accurate than other tools. While previous reviews have examined diet assessment tools for children [25][26][27] few have included preschool-aged children, have restricted inclusion of assessments to short forms (i.e., ≤50 items) [28] or have narrowly focused on interventions aimed at the physical home food environment without social characteristics. Thus, the objectives of this review were to: (1) describe the dietary assessment tools used in intervention studies in young children focused within the home environment; and (2) examine how the application of these dietary assessment tools addresses intervention objectives.

Data Sources and Searching Strategies
The protocol of this review was developed based on the current recommendations set by the Preferred Reporting Items for Systematic Reviews and Meta-analysis Approach (PRISMA). Two electronic databases, PubMed and Web of Science, were searched in February 2022. The searching strategy was developed using a combination of Medical Subject Headings (MeSH) terms and key words relating to young children, home-environment interventions, diet assessment tools, intake and study type. University librarians, with expertise in systematic reviews, provided guidance on search terms to optimize the search strategy created for PubMed and adapt it for use in Web of Science. A time filter was used in both searches to include articles published only after 1 January 2010, which was selected to capture the most recent evidence and minimize overlap with studies included in prior reviews. The full list of searching terms is in Appendix A. In addition to the database search, manuscripts were identified via review of reference lists as well as studies citied within the selected manuscripts, particularly when a secondary analysis or methodology paper was initially identified.

Inclusion and Exclusion Criteria
To be included, manuscripts must have been written in English and published since 2010. Studies must have also included only healthy, typically developing preschool-aged children, 2-5 years, without known disorders or conditions that would have the potential to impact dietary intake (e.g., severe allergy or cystic fibrosis). Included studies must have conducted a nutrition intervention with a component focused on the home environment, including both the physical (e.g., food availability) and social (e.g., parent feeding practices) environments. Additionally, studies must have measured the child's solid food and/or beverage intake (not infant formulas or breastmilk) in the home environment using a diet assessment tool and reported on a diet related outcome(s).
Secondary analysis, methodology manuscripts, review papers, cross-sectional studies, and validation studies were excluded from this review. Additionally, studies were excluded if the majority (>50%) of children fell outside 2-to 5-year age range at baseline, or if the study population was primarily on the upper or lower ends of the age range (e.g., 5-7 years old or 0-2 years old).

Selection of Studies
Following the search of databases, the identified manuscripts were imported to systematic review software, COVIDence (Veritas Health Innovation, Melbourne, Australia). After the removal of duplicates, title and abstracts were screened independently by two reviewers to identify eligible studies according to the inclusion and exclusion criteria. Full text articles of manuscripts deemed eligible were then imported to COVIDence for fulltext review. The two reviewers independently conducted a full text review by including manuscripts that met inclusion criteria, then eliminated and recorded the reason(s) for each excluded manuscript. Any conflicts during the title and abstract review or full-text review were resolved through discussion between reviewers, and if needed, the additional consultation of a third reviewer. After completing the full-text review, included manuscripts that were written about the same intervention program were identified, and the most relevant manuscript was selected. This way, each intervention, rather than each manuscript, was the unit of interest for the review. For this review, manuscripts reporting study outcomes were prioritized.

Data Extraction
A template was created in Excel to extract study characteristics and outcome data. Data extraction was split and completed by two reviewers independently. For quality assurance, a subsample of included articles (20%) was completed by the two reviewers together. The following information was extracted: bibliographic information (first author's last name, year of publication, and country of study), study population (children's age, sample size, adult respondents' role/gender, race, education level, and income level), interventionrelated (intervention name, objective, target, length, and diet-related components in the intervention), and diet assessment-related (diet assessment tool(s), description of the tool(s) used, outcome variables, and key findings).

Results
A PRISMA flow chart that summarizes the results of the search, including reasons for exclusion is shown in Figure 1. The database searches initially identified 370 articles. After duplicates were removed, 297 articles were moved into titles and abstracts screening, and 58 were accessed and reviewed in full text for eligibility. During full text review, 5 manuscripts outside the initial database searches were identified from the references and subsequently imported manually for review. A total of 17 studies were included in this review.
Nutrients 2022, 14, x FOR PEER REVIEW 6 of 21 the most relevant manuscript was selected. This way, each intervention, rather than each manuscript, was the unit of interest for the review. For this review, manuscripts reporting study outcomes were prioritized.

Data Extraction
A template was created in Excel to extract study characteristics and outcome data. Data extraction was split and completed by two reviewers independently. For quality assurance, a subsample of included articles (20%) was completed by the two reviewers together. The following information was extracted: bibliographic information (first author's last name, year of publication, and country of study), study population (children's age, sample size, adult respondents' role/gender, race, education level, and income level), intervention-related (intervention name, objective, target, length, and diet-related components in the intervention), and diet assessment-related (diet assessment tool(s), description of the tool(s) used, outcome variables, and key findings).

Results
A PRISMA flow chart that summarizes the results of the search, including reasons for exclusion is shown in Figure 1. The database searches initially identified 370 articles. After duplicates were removed, 297 articles were moved into titles and abstracts screening, and 58 were accessed and reviewed in full text for eligibility. During full text review, 5 manuscripts outside the initial database searches were identified from the references and subsequently imported manually for review. A total of 17 studies were included in this review.
The sampling period varied across the different tools. Of the studies that used an FFQ or FFQ Screener, 6 studies had a sampling period that ranged from the previous 1 day to 6 months [25,28,29,31,32,37], and 3 studies did not report the sampling period [39,42,44]. For studies that used a Food Record, the sampling period was 3 days in 3 studies [36,39,43] and 4 days in 1 study [44]. All 4 studies included 1 weekend day. Studies that used a 24 h Recall had a sampling period of 2 days in 1 study [32] and 3 days in 2 studies [31,33], and all included 1 weekend day. Of the 3 studies that used a Food Checklist, the sampling period for 1 study was for the previous 1 week [45], and the remaining 2 studies did not report the sampling period [30,35].

Intervention Scope
Interventions ranged in length from 3 weeks [44] to 3 years [31] and focused on a variety of topics related to the physical and social home food environment. Eleven interventions focused primarily on the social home food environment and addressed topics related to feeding practices, behavioral capability (knowledge and skills), self-efficacy, and role modeling [31][32][33][34][35]37,38,42,43,45,46]. Two studies focused on these social topics as well as the physical home food environment by addressing availability and accessibility of healthy or unhealthy foods with parents [40,41]. Two studies focused primarily on the physical home food environment by making target foods available to children [39,44]. Two studies provided limited information on intervention content [30,36].

Intervention Objectives
Diet-related intervention objectives, as described in the studies, were categorized into 4 groupings: 7 studies focused on overall diet [35,37,38,41,43,45,46], 4 on specific food groups [30,33,39,40], 1 study on eating occasions [44], and 5 studies on obesity prevention or treatment [31,32,34,36,42]. Among the studies grouped as overall diet, the primary objectives explicitly listed a dietary objective yet varied in terminology-dietary diversity [45], dietary intake [37], dietary patterns [38], food consumption [41], nutritional status [43], and nutrition [35,46]. Studies on specific food groups had objectives centered on fruits and/or vegetables [30,39,40], solid fats and added sugars (SoFAS) [33], or junk foods [35]. The only eating occasion study had objectives on snacks [44]. Among the 5 studies grouped in the obesity prevention and treatment category, all studies had broad objectives and only 2 listed diet or nutrition in the objectives [32,36]. None of the 5 studies listed specific diet-related study objectives yet all reported findings related to diet outcomes. Table 3 presents intervention objectives, diet assessment measures and key findings for each study.

Key Findings
Each study was reviewed to assess concordance of (1) diet-related intervention objectives with key findings, (2) key findings with the diet assessment tool used, and (3) key findings produced from multiple tools within individual studies. Eight of the 17 studies reported key findings that aligned with their objective and assessment tool(s) [30,33,36,[39][40][41]44,45]. Key findings from 1 study were discordant with both the intervention objective and tool used, where its objective was on dietary patterns, its findings were on energy, and its assessment tool was an FFQ, which is suboptimal for measuring energy intake [38]. Three studies were partially aligned, where 1 matched in objectives and findings but used an FFQ to capture energy [37]; 1 inquired on nutrition and reported on dietary diversity but used an FFQ without baseline data [46]; and 1 inquired on nutritional status but reported on consumption patterns despite using an adequate tool (i.e., Food Record) for its key findings [43].
Five of the studies had objectives that were broad or did not specify diet in their objective yet reported on specific dietary outcomes [31,32,34,35,42]. Among those that were broad, 1 study broadly inquired about nutrition and reported specifically on fruit, vegetable, and junk food consumption [35]; another inquired about food parenting and weight outcomes and reported broadly on macronutrients and energy [32]. Those that did not specify diet in their objectives included one that inquired about Body Mass Index and reported on energy [31], another that inquired about healthy behaviors and reported on fruit and vegetable intake [34], and a third that inquired about obesity treatment and reported on sweet consumption [42]. Four of the five studies with objectives that were broad or did not specify diet-related outcomes were in the obesity prevention and treatment grouping [31,32,34,42].
Three of the 17 studies used multiple dietary assessment tools, 2 using a mix of self-report and objective measures to answer the same question [30,39] and 1 using multiple self-report measures to answer different research questions [44]. Bakirci-Taylor et al. (2019) [30] used 3 different tools: 1 self-report (Food Checklist) and 2 objective measures (Food Photography and Veggie Meter). For intervention effects, findings were discordant between data from self-report and objective tools, where the self-report showed a positive intervention effect and objective showed partial or no effect. Poelman et al. (2019) [39] used 3 different tools, 2 self-report (FFQ and Food Record) and 1 objective (Weighed Dinner Meal). For intervention effect, findings were discordant as FFQ showed partial positive intervention effect, Food Record showed no effect, and Weighed Dinner Meal showed no effect. Reale et al. (2018) [44] used 2 self-report tools (FFQ and Food Record). For intervention effect, differences were detected between snack control strategies at increasing vegetable intake (Food Record) and no changes in frequency of snack intake was detected (FFQ). Evaluate the effectiveness of family participation in nutrition education on the nutritional status of preschool-aged children. Determine if provision of quality nutrition information to rural parents using self-directed, technology-based education resources has an effect on the dietary patterns of 2-5-year-old children.

M FFQ (Previous 6 M) B, 3 M, M
Total reported energy from nutrient-dense food groups and percentage energy from energy-dense, nutrient-poor foods were high at baseline relative to estimated total energy expenditure for child age. No intervention effect.

Natale et al. (2014), USA [35]
Assess the effectiveness of a child care-based parent and teacher healthy lifestyle role-modeling program on child nutrition and PA outcomes. Evaluate the effects of a preschool-based family intervention on children's energy balance-related behaviors such as food consumption and screen time, as well as PA and self-regulation skills.

M FFQ (Previous 1 W) B, M
No significant differences were detected between intervention and control groups for energy balance-related behaviors, self-regulation skills, or for consumption frequencies of sugary everyday foods and beverages, sugary treats, and FV. Evaluate the efficacy of an authoritative food parenting intervention for low-income mothers to reduce preschool-aged children's intake of calories from SoFAS.

W 24 h Recall (3 d) B, W
At post-intervention, children in the intervention group consumed~94 kcal or 23% less daily energy from SoFAS than control group. Child total daily energy intake did not significantly differ between groups post-intervention. Assess the efficacy of a telephone-based intervention for parents to increase FV consumption in their 3-5-year-old children. Explore the preliminary efficacy of two strategies of snack portion control (reduction and replacement) and; examine the efficacy of these two methods to improve diet in preschoolers.

W
FFQ, (NR) 1 W, 7-9 W Snack replacement resulted in increased vegetable intake and decreased total daily energy intake when compared to snack reduction (Food Record). No significant changes to the frequency of snack intake pre-and post-intervention (FFQ). Food Record (4 d) 1 W, 2 W, 3 W

Obesity Prevention and Treatment
Barkin et al. (2018), USA [31] Test the effect of a multicomponent behavioral intervention on child BMI growth trajectories among preschool children at risk for obesity.

M 24 h Recall (3 d) B, 12 M, 24 M, M
The intervention resulted in a statistically significant reduction in mean child daily energy intake and higher percentage of energy from protein compared with control. Davison et al. (2013), USA [32] Test initial efficacy of a parent-centered, community-based participatory research approach for obesity prevention for improving food, PA, and media-related parenting, as well as child behavioral and weight outcomes.

M 24 h Recall (2 d) B, 8 M
At post-intervention, children had significantly lower total energy intake and macronutrient intake (fat, protein, and carbohydrate) compared with pre-intervention. Examine the feasibility and preliminary efficacy of using Facebook in a lifestyle intervention to improve healthy behaviors and reduce BMI. Examine the effectiveness of a home-based obesity prevention intervention on health behaviors, obesity risk and dietary intakes.

M Food Record (3 d) B, M
Post-intervention, the 4 home visits (HV) group had significantly higher fiber intake. The 4 HV and 2 HV groups had significantly higher fruit intake compared to control.

Somaraki et al. (2020), Sweden [42]
Evaluate the effects of two approaches (a parent support program with and without booster sessions and standard treatment) to treat obesity in preschoolers.

Discussion
This review summarized the application of dietary assessment tools used in intervention studies focused on young children in the home environment. Of the 17 studies examined, 12 studies used fixed question sets to examine dietary intake over a designated time frame (9 FFQ [34,(37)(38)(39)(40)(41)(42)44,46] and 3 food checklists [30,35,45]) and 7 used open-ended assessments over multiple 24 h periods (3, 24 h recalls [31][32][33] and 4 food records [36,39,43,44]). Only 6 studies reported on the psychometric properties of tools that were tested with preschool audiences [32,35,37,38,41,42]. The studies in this review had intervention objectives that were grouped into 4 categories: overall diet, foods/food groups, eating occasion, and obesity prevention and treatment. Intervention objectives varied in specificity and the use of nutrition terminology. Only 8 of the studies' intervention objectives fully aligned with their study findings [30,33,36,[39][40][41]44,45]. An additional 3 studies [37,43,46] partially aligned and 1 study was discordant [38]. Five studies, four of which were in the obesity prevention and treatment grouping, were broad or did not have a specific dietary objective [31,32,34,35,42]. Parents and caregivers served as proxy respondents in all studies, with 2 studies collecting complementary data directly with children through objective measures-weighed dinner [39] and Veggie Meter [30].
Findings suggest that dietary assessment tools used with preschool-aged children are most often applied to answer a variety of research questions with no agreement on standards for assessment, including which foods to measure, the time frame to consider, and related social characteristics impacting the home food environment. Further, some but not all intervention objectives matched reported key findings. Collectively, these findings point to challenges in obtaining dietary data on preschool-aged children, particularly with the reliance on parents and caregivers as proxy respondents, as well in interpretation of the data obtained from the self-report dietary measures.
To improve children's diet, it is necessary to accurately measure their current intake. The accuracy is reliant on validity of the tool itself as well as reliability of the respondents' input. Reporting the dietary intakes of young children, particularly in the context of obesity, brings with it additional challenges and considerations. These include the need for a proxy respondent (e.g., parent or caregiver), consideration of developmental stage (e.g., cognitive skills), and food consumption away from home [18]. All studies in this review used parents or caregivers as respondents, with 7 studies using a 24 h recall or daily food records [31][32][33]36,39,43,44]. The likelihood of recall and reporting bias are high for proxy respondents, particularly if they are asked to report on 24 h intake in which they are not present for some of the eating occasions or if multiple respondents split the reporting over the 24 h period [47]. Thus, researchers should consider if a full day's intake is necessary versus examining eating occasions in which the caregiver is with the child (e.g., dinner) and/or those time intervals in which the intervention is specifically targeting. If the intervention target is on parents serving their children more fruits and vegetables, then collecting dietary data when the child is outside of parental care may not be reliable from a measurement perspective, nor produce a valid intervention effect. Lastly, respondent characteristics such as educational level, reading and digital literacy, and inclusion of food relevant to respondents' culture (e.g., ethnic background, geographical location) should be considered for accuracy along with respondent burden.
The accuracy of dietary data relies on the use of psychometrically sound tools that are critically assessed with the sample population. The reporting of psychometric properties for the studies included in this review were limited in detail and scope. Nine studies in this review did not fully report psychometric properties of reliability and validity [30,32,[34][35][36][37][38][39]41,42] and an additional 3 studies reported sources in which testing was done on populations other than preschool-aged children [30,34,39]. This illustrates not only the need for better measures for early childhood audiences, but also more awareness and/or training amongst researchers in dietary measurement. This is critical as the use of methods with low validity greatly attenuates the associations between dietary intakes and outcomes in health [16].
Due to the inherent limitations of self-report diet assessment tools, combined with challenges of working with early childhood audiences and sparse psychometrics, considerations for using multiple methods to maximize the strengths of each instrument may be warranted. The use of alternative approaches to capture diet as well as interpret outputs from dietary measures may provide researchers the ability to answer more clearly defined research questions to strengthen the concordance of intervention objectives and study design. Two studies in this review coupled self-report dietary data with objectives measures, including food photography and biomarkers (e.g., Veggie Meter) [30] and a weighed dinner meal [39] to evaluate different aspects of their intervention. A third study used multiple dietary tools to examine different aspects of their intervention related to an eating occasion, snacking [44].
As technology becomes more ubiquitous with daily life, utilizing digital devices to capture dietary information in young children has potential. Food photography is an emerging tool for dietary assessment as an image-assisted method to enhance another dietary assessment method, as used in the Bakirci-Taylor et al. (2019) study [30], or as the primary form of dietary data [48]. Food photography has been used with preschool audiences to assess a 24 h period [49] or a single eating occasion, such as lunch [50] and dinner [51,52]. In addition to providing energy and nutrients [52], data produced by food photography has been used to examine dietary quality via the Healthy Meal Index [51,53] and comparison to dietary recommendations, namely the Child and Adult Care Food Program's nutritional standards [50,52]. Lastly, McCloskey et al. (2019) [51] used food photography to examine the context of foods served to preschoolers in the home environment by examining concordance with mothers' foods served and timing of meal, as well as food preparation (e.g., takeout to non-convenience). Food photography provides an alternative method to collecting dietary data for preschool audiences as well as innovative approaches to interpreting these data.
In addition to alternative methods, researchers should consider alternative approaches to capturing food or diet-related data beyond just intake. When examining studies in this review, only 2 studies [40,41] addressed food availability and accessibility in their interventions, with an additional 2 studies [39,44] making target foods available to their participants. Both cross-sectional and longitudinal studies demonstrate that home food availability can predict child food intake [10,54,55]. For instance, a study by Boles et al. (2019) [10] examined the relationship between the HFE and child dietary intake of preschoolaged children from rural and low-income, culturally diverse families and found that the availability of fruits and vegetables, meat products and sugar sweetened beverages were shown to predict greater intake by preschool children [10]. Because food availability has been shown to be predictive of dietary intake, researchers may consider using measurement tools that capture the availability of foods in the home as an alternative or complement to dietary intake measures.
There are several limitations to this review. First, it only included peer-reviewed full-text, English publications, hence publications in languages other than English or in the gray literature may exist and provide additional insights on this topic. Next, we only included 2 databases in our review. Collectively, these limitations may have influenced the authors' ability to find and interpret the literature on this topic. This narrative review has several strengths, including using the most updated guidelines on completion of systematic reviews (PRISMA). Two reviewers screened full articles for inclusion and extraction with a high level of agreement. To note, while neither a strength or limitation, most of the intervention studies included in this review focused on childhood obesity and reported diet as a primary outcome. This is likely due to the high global prevalence, resulting in increased funding for such studies during the time period of our search (2010-2022).

Conclusions
This review found variation in the concordance of key findings with intervention objectives, type of dietary assessment tool used, and multiple tools within the same study. The use of traditional, self-report dietary assessment measures with preschool audiences to capture accurate data is challenging. The need for proxy respondents, coupled with limited psychometric properties of existing tools, calls for more methodological work as well as innovative approaches to capture and interpret young children's dietary data. Approaches to advance objective measures, such as food photography, and the expansion of research questions beyond just daily intake are warranted. This could include examining food availability in the home environment or capturing intake by eating occasions instead of over a full day in which the child is in various settings (e.g., preschool, home) with multiple respondents (e.g., parent, teacher, another caregiver). These types of approaches could improve the concordance between intervention objectives, dietary assessment tools, and key findings, and ultimately provide a more robust understanding of young children's diets and the efficacy of nutrition interventions.  Appendix A.2. Web of Science Search Terms TS = ("young child*" OR kid OR preschool* OR toddler*) AND TS = ("parent" OR "home environment" OR "home-based" OR "home food environment" OR "family meal" OR "home food" OR "food in home" OR "family-based" OR "family environment" OR "family food environment" OR caregiver OR parent OR famil* OR mother OR father OR guardian) AND TS = ("diet recall*" OR "food recall*" OR "24 h recall*" OR "food frequency questionnaire*" OR FFQ OR "screener" OR "dietary questionnaire" OR "food survey*" OR "diet survey*" OR "meal survey*" OR "food record*" OR "diet record*" OR "food checklist*" or "food behavior checklist*" OR "diet checklist*" OR "food photo*" OR "photo assisted diet assessment*" OR PADA OR "image based food record*" OR IBFR OR "dietary observ*" OR "meal observ*" or "food observ*" OR "diet assess*" OR "food assess*" OR "meal assess*" OR "weighed record*" OR "plate waste" OR "direct weighing method*" OR "food diar*" OR "meal diar*") AND TS = (Interven* OR trial OR RCT OR "randomized control trial" OR "randomized control trial" OR "quasi experiment*" OR "behavior change")